A fault identification scheme based on integrated wavelet neural networks was proposed to deal with the fault identification problem of the satellite power systems.Firstly,the critical data of the satellite power system was utilized to identify a certain class of failure group preliminarily.Secondly,characteristics of this type of fault group were studied,and corresponding wavelet neural network was constructed for the further fault identification.Finally,simulation for a class of satellite power system was conducted,and the result verified the effectiveness of the proposed method.