Aiming at the difficult problems of complex types of faults, difficulties in obtaining typical fault information and being subject to the impacts of environment, such as noise and temperature in analog circuit fault diagnosis, an analog circuit fault diagnosis method based on random projection and naive Bayesian network is proposed in this paper. The method firstly extracts the analog circuit fault information, and random projection method is used to perform the dimension reduction of the fault features and obtain the analog circuit fault feature vectors. Then, the naive Bayesian classifier diagnostic model is used to identify various faults of the analog circuit. The experiment results of CSTV filter circuit, four opamp biquad highpass filter circuit and actual SallenKey bandpass filter circuit indicate that compared with traditional analog circuit fault diagnosis method, the proposed method shows higher fault diagnosis performance and stronger antiinterference ability.