A new method of Vibration hydro-generating Unit fault diagnosis was proposed base on Choquet Integrate combined classifier for improve fault diagnosis accuracy, improve Naive Bayes Classifier base on Attribute Similarity Weighted Naive Bayes Classifier(SWNBC),use Mahalanobis Distance Classifier(MDC) and BP neural network (BP) composed combined classifier SWNBC+MDC+BP, wavelet packet extraction of related band energy as input characteristic vector.Applicate combined classifier for fault diagnosis, fuzzy integral to make the final fault type. The simulation experiment shows that the model is better than the single classifier, can effectively identify fault.