The N-type insulated gate bipolar transistor (N-IGBT) is widely used in various fields of modern industries due to its excellent performance. The prediction of the device degradations under specific conditions is of great significance for improving N-IGBT reliability. However, with the decrease of the N-IGBT process, the internal electric field intensity of the gate oxide is constantly increasing, and the degradation of the gate oxide caused by the positive bias temperature instability ( PBTI) is further intensified, the degradation is macroscopically reflected in the reduction of device remaining useful life (RUL) and changes of the threshold voltage. Based on the classical Power-Law model and the Arrhenius model, this article proposes a three-stage Power Law-Arrhenius comprehensive degradation model with relatively higher accuracy using the common parameter degradation time as the starting point. The degradation effect of positive bias temperature instability on N-IGBT through accelerated degradation experiment. Then, the threshold voltage which reflects the life of the power device after degradation is measured. Based on the genetic optimization algorithm and accelerated degradation experimental data, the comprehensive degradation model parameters are fitted and optimized, the mathematical expression form of the comprehensive degradation model is determined with an its accuracy above 85% , which is higher than that of the traditional power law model.