Abstract: This paper took the open circuit fault in Insulated Gate Bipolar Transistor (IGBT) dual transistors in CRH3C EMU inverter as the research object, proposed an inverter fault diagnosis method based on the African Vultures Optimization Algorithm (AVOA) and an optimized Back Propagation (BP) neural network. The paper built a control model for train inverters in Simulink and obtained faulty electric current, used Empirical Mode Decomposition (EMD) to denoise and extract fault features from the electric current signal, and then used AVOA to optimize the BP neural network, implemented the diagnosis of open circuit faults in IGBT dual transistors of train inverters. Compared with traditional methods, this method has higher accuracy with an accuracy of 100% in the test set.