In response to the problems of low knowledge utilization rate, single reasoning method and fixed knowledge base in traditional fault diagnosis expert systems, a fusion reasoning expert system based on fuzzy fault tree is proposed. Firstly, a fuzzy fault tree for belt conveyor was established, and the qualitative and quantitative analysis of faults was realized by combining the intuitionistic fuzzy algorithm, forming a standardized knowledge base and case base for the expert system; Secondly, a knowledge case fusion reasoning algorithm was designed as the reasoning engine of the expert system, enabling it to realize real-time monitoring and reasoning of faults by combining the state monitoring module; Finally, a knowledge base dynamic update algorithm based on BERT model was developed, converting case data actively into knowledge rules to improve the operation and maintenance efficiency of the system. The performance of the expert system was tested through experiments, and the experimental results show that: the average diagnostic accuracy of the proposed expert system can reach 94.75%, and the average reasoning delay is 167.31ms; The designed knowledge base dynamic update algorithm significantly improves the operation and maintenance manpower savings in large case library environments; The key performance parameters of the system increase with the increase of the number of cases, and the expert system shows good scalability and growth potential.