The expert system, one of the artificial intelligence research field, is widely used to meet the need of engineering equipment. However, the examples of satisfying the requirements of users are quite scarce due to the complexity of establishing the knowledge base and the lack of the guarantee of the quality. To solve this problem, the problems of knowledge acquisition were studied and the knowledge base building tool software based on fault tree was developed, which realized the function of fault diagnosis. According to the actual demand, the J2EE technique was used to develop a set of tool software for establishing knowledge base on B/S (browser/server) model, and the requirement analysis design of each module of knowledge base were carried out, which involved the design of model data structure and the design of business layer logical method. In addition, the diversity expression of knowledge model was discussed, and the complete fault tree was expressed in three naming ways. Finally, an example was given to illustrate the feasibility of the knowledge base building tool software. The results indicated that the expression way of fault tree was used as the core of knowledge-acquiring module to improve the quality of knowledge base. In the meanwhile, the Web form was selected to realize editing and inputting knowledge in multi-user/multi-workstation model to raise the efficiency of acquiring knowledge. Thus, this auxiliary system for knowledge acquirement has strong universality, and it can provide powerful support for domain experts and engineers to establish knowledge base.