About Resources Disclaimer rd (at) patentbusinesslawyer (dot) com
Software patents · China

Software Patents in China

China’s current patent examination framework can protect software-related inventions when the claimed solution uses technical means to solve a technical problem and obtain a technical effect. CNIPA’s 2026 examination amendments reinforce whole-claim analysis for inventions containing algorithmic or business-rule features.

Save or follow this source
Software patent strategy in China
Software × Chinaeligibility · claims · prosecution · PCT

Software patent decision framework

Primary intentSoftware patentability, claim drafting and examination strategy in China.
Procedural filing routeUse the dedicated national/regional phase page for procedural entry requirements.
International parentInternational software patent filing strategy.
Related decisionsPCT, where to file, cost, timeline.

Software Patents in China — researched guide

This page owns the software-specific patentability and prosecution question for China. Filing deadlines, translations and formal national-phase requirements remain on the dedicated procedural page.

Current CNIPA guidance should control the analysis

CNIPA’s 2025 amendment to the Patent Examination Guidelines took effect on 1 January 2026. The revised provisions address inventions involving AI, big data, algorithm features and business rules. They require examination of the claimed solution as a whole rather than mechanically separating technical and algorithmic features. Software applicants should therefore analyze the interaction between program logic and the technical system instead of assuming that the presence of an algorithm automatically excludes protection.

Technical problem, technical means and technical effect remain central

The revised Guidelines state that a solution containing algorithm or business-rule features can qualify as a technical solution where it uses technical means to solve a technical problem and obtain a technical effect. Conversely, a claim that does not solve a technical problem, use technical means or obtain a technical effect may fall outside patentable subject matter. The specification should make each element of this technical chain explicit.

Algorithmic features can contribute to inventive step

CNIPA’s current guidance recognizes that algorithm or business-rule features can contribute to inventive step when they functionally support and interact with technical features. This makes the relationship between software logic and technical implementation important. A claim should not present the algorithm as a detached mathematical process if its commercial value actually arises from how it changes data acquisition, device operation, network processing or another technical function.

Disclosure should explain how the software is implemented

For software-related inventions, the specification should describe the solution clearly and completely enough for the skilled person to implement it. Where algorithms are important, applicants should explain the necessary modules, processing relationships, input and output data and the role of the algorithm in the technical scenario. High-level product descriptions may leave insufficient support for later amendments.

Software claims should be structured around controllable technical acts

China-facing software applications should support claims directed to systems, methods, devices and other permitted forms where appropriate. Distributed systems should identify which technical steps are performed by the server, client, terminal or other component. This improves prosecution flexibility and can later affect infringement analysis.

Business-software inventions need stronger technical framing

Fintech, commerce, logistics and platform software frequently include business rules. The CNIPA framework does not make a commercial rule patentable simply because it runs on a computer. The application should isolate any technical contribution, such as improved network communication, device authentication, resource allocation, data processing or security, and explain the interaction between those technical features and the software logic.

AI and software strategy increasingly converge

The 2026 CNIPA amendments expressly address AI, big data and algorithms. Software companies using AI should consider whether model construction, training, data processing or application-specific technical use forms part of the invention. The disclosure should explain the model’s technical role rather than relying on generalized AI terminology.

China supply chains can change filing value

For software embedded in devices, industrial equipment, telecom products or manufacturing systems, China may be important even when the applicant’s largest customers are elsewhere. Filing strategy should consider manufacturing, integration, local deployment, competitors and licensing partners. Pure SaaS businesses may have a different enforcement and infrastructure map.

PCT entry should preserve technical support

The PCT application should already contain the technical detail needed for CNIPA examination because later national-phase amendments cannot simply introduce new matter. Applicants should preserve Chinese claim options in the international specification, including alternative technical implementations and sufficient examples. Procedural national-phase requirements belong on the dedicated China PCT page.

Decision checklist for China software patents

Before filing in China, review the technical problem, technical means, technical effect, interaction between algorithmic and technical features, disclosure depth, claim architecture, AI overlap, business-rule content, Chinese commercial or supply-chain relevance, PCT support and infringement evidence. The strongest applications connect the software logic directly to a defined technical solution.

How to prioritize software claims for China

A practical filing strategy should rank claim sets by commercial importance and enforcement value instead of treating every disclosed feature as equally important. The first layer should capture the technical function that differentiates the product. A second layer can protect narrower architectural details, data structures, security mechanisms, resource-management techniques or device interactions that may be easier to defend during examination. A third layer can address deployment-specific embodiments, such as server-side, client-side, edge or embedded implementations. This layered approach matters because prosecution can require amendment, and the original specification must already support those fallback positions. The goal is to preserve meaningful commercial scope while giving the examiner technically concrete alternatives.

Evidence of infringement should influence drafting

Software patents can lose practical value when the patented function occurs entirely inside a competitor's inaccessible backend. Before filing, the applicant should ask what evidence could show that the claimed steps are actually performed. Public API documentation, product behavior, network messages, device outputs, customer manuals, standards documents and observable system responses can all matter. Where an important feature is hidden, claims should, where the disclosure permits, include externally observable consequences or system relationships. This is separate from patentability, but it directly affects portfolio value. A technically valid patent that cannot be mapped to a competitor's product may provide less leverage than a narrower claim tied to observable operation.

Portfolio review before major foreign-filing costs

Software products change quickly, so the original filing map should be reviewed before PCT national or regional phase entry, examination requests, grant and renewal. The company should confirm that the protected feature still appears in the product roadmap, that competitors remain active in the target jurisdiction, and that the local market supports licensing, enforcement or defensive value. The patentability analysis should also be refreshed against current patent-office practice before committing substantial prosecution cost. This review can lead to continuation, divisional or follow-on filings where the product has evolved, or to abandonment where the commercial case has weakened. Portfolio discipline is especially important for software because product cycles are often shorter than patent lifecycles.

Related specialist resources: For patent analytics and evidence work, use GIP Research patent analytics and GIP Research open research. For commercialization and transaction strategy, see Patent Business Lawyer and Patent Business Attorney. For adjacent technology-law questions, use US technology-law guidance, TechCorpLegal corporate technology-law resources, and Dr. Rahul Dev research and professional resources.
Related mobile-specific guide: Mobile app patents in China.
Featured insight

Software claims should be drafted for the jurisdiction and the real deployment architecture.

Eligibility, technical effect, actor structure, cloud deployment and infringement evidence should be considered before national or regional phase costs are incurred.

Featured Work & Insights

Discuss an international software patent strategy

rd (at) patentbusinesslawyer (dot) com

Email Dr. Rahul Dev

How to prioritize software claims for China

A practical filing strategy should rank claim sets by commercial importance and enforcement value instead of treating every disclosed feature as equally important. The first layer should capture the technical function that differentiates the product. A second layer can protect narrower architectural details, data structures, security mechanisms, resource-management techniques or device interactions that may be easier to defend during examination. A third layer can address deployment-specific embodiments, such as server-side, client-side, edge or embedded implementations. This layered approach is useful in China because prosecution can require amendment, and the original specification must already support those fallback positions. The goal is to preserve meaningful commercial scope while giving the examiner technically concrete alternatives.

Evidence of infringement should influence drafting

Software patents can lose practical value when the patented function occurs entirely inside a competitor's inaccessible backend. Before filing, the applicant should ask what evidence could show that the claimed steps are actually performed. Public API documentation, product behavior, network messages, device outputs, customer manuals, standards documents and observable system responses can all matter. Where an important feature is hidden, claims should, where the disclosure permits, include externally observable consequences or system relationships. This is separate from patentability under technical problem, technical means and technical effect, but it directly affects portfolio value. A technically valid patent that cannot be mapped to a competitor's product may provide less leverage than a narrower claim tied to observable operation.

Portfolio review before major foreign-filing costs

Software products change quickly, so the original filing map should be reviewed before PCT national or regional phase entry, examination requests, grant and renewal. The company should confirm that the protected feature still appears in the product roadmap, that competitors remain active in the target jurisdiction, and that the local market supports licensing, enforcement or defensive value. In China, the patentability analysis should also be refreshed against current CNIPA practice before committing substantial prosecution cost. This review can lead to continuation, divisional or follow-on filings where the product has evolved, or to abandonment where the commercial case has weakened. Portfolio discipline is especially important for software because product cycles are often shorter than patent lifecycles.