Aiming at the traditional fault diagnosis method of neutral point clamped inverter(NPC) based on signal analysis, it is prone to spectrum leakage and has the problems of signal nonlinearity and unsteady state. An inverter open-circuit fault diagnosis method based on the Prony algorithm is proposed. Firstly, wavelet transform is used to denoise the original measurement signal to overcome the disadvantage that the Prony method is sensitive to noise. Secondly, the Prony analysis method is used to extract the four-dimensional fault characteristics of amplitude, frequency, phase, and attenuation factor by fitting the three-phase current signal of the inverter. Then, a support vector machine (SVM) is used to classify the neutral point clamped inverter faults. Finally, the proposed algorithm is verified on the simulation platform. The simulation results show that the transient nonlinear current signal of neutral point clamped inverter can effectively extract the fault characteristics and realize the open circuit fault diagnosis of single and double transistors, and this method is fast and accurate in diagnosis.