目的实现运载火箭故障诊断的智能化、自动化和实时性，提高运载火箭发射的成功率．方法引入人工智能技术，研制故障诊断专家系统．结果原型系统现有知识点２１０００余个，其中规则２０００余条．系统中最大的一项诊断时间为０６７ｓ，最小为０１３ｓ，小于平均每一项的诊断时间不超过１００ｓ的需求．结论专家系统用于火箭故障诊断具有可行性和实用性． Aim To realize intelligence, automation and real-time of fault diagnosis of launch vehicle and to improve the success rate of launch of launch vehicle. Methods The introduction of artificial intelligence technology, the development of fault diagnosis expert system. The results of the prototype system has more than 21,000 knowledge points, of which more than 2000 rules. The largest diagnostic time in the system is 067s, the minimum is 013s, less than the average of 100s for each diagnosis time. Conclusion The expert system is feasible and practical for rocket diagnosis.