Safety and reliability of power batteries have become the key factors restricting the development of electric vehicles. By using fault tree and FMEA analysis, the common faults of power battery are classified and summarized. Using deep belief network and decision fusion algorithm, a power battery fault diagnosis model based on deep learning is established. The model establishes the DBN network model through the performance parameters of power battery samples, calculates the support and confidence of faults, and combines the decision fusion algorithm to obtain a more accurate fault fusion diagnosis result. BYD E5 power battery fault parameters are used for model data simulation, which validates the effectiveness of the fusion algorithm. The results show that the model has high accuracy and stability for fault classification and diagnosis of power batteries.

CICTP