In this paper, an open circuit fault diagnosis method for inverter based on echo state network (FD-ESN) is proposed. Firstly, the voltage law at both ends of load is analyzed under normal operation and single/two insulated gate bipolar transistors (IGBTs) open circuit faults, and the input and output samples of the prediction model are collected. Secondly, a fault prediction model based on echo state network (ESN) was built and its weight parameters were optimized. Finally, the trained ESN fault prediction model is used to diagnose the inverter. The simulation results show that the proposed FD-ESN method can accurately predict the open-circuit faults of the inverter in real time.