知识库系统是一种专门存储、管理大量知识的机构。文章以可拓学的理论和方法为指导，提出可拓知识表示法和可拓知识库系统的特点与构架。给出了可拓知识表示的形式化语义，其基本单元是知识物元；阐述了矛盾问题求解的形式表示，其核心是问题的物元模型，通过可拓算子把不相容问题转化为相容问题；相关方式推理是可拓知识库系统推理机的特点，是传统规则推理的推广。可拓知识库采用可拓知识表示方法，是传统知识库的一个超集；可拓知识库系统由目标库、条件库、公用知识库、分类知识库、推理机与知识库管理系统组成。

Knowledge base system (KBS) is an organization to store and manage a lot of knowledge. Extension knowledge representation (EKR) based on extension theory is presented in this paper. Two key concepts of EKR are the described——formal semantic of EKR and the formal representation of contradictory problem solving. The basic element of the former is knowledge matter-element model. In addition, the Specific Extension Inference (SEI) corresponding to EKR is also reviewed. The feature of SEI is correlative inference, which has been implemented by computer. Finally, the feature and the frame of extension knowledge base system (EKBS) are proposed. EKBS uses EKR and is an excess of traditional KBS. It is composed of goal system, condition system, common knowledge base, categorized knowledge base, inference machine and knowledge base management system.