The traditional signal spectrum analysis method is applied to the diagnosis of stator interturn short circuit fault with low signaltonoise ratio and with no obvious fault characteristics, which often leads to the problem of misjudgment. Therefore, a stator interturn short circuit fault diagnosis method based on highorder spectrum is proposed. Compared with the traditional spectrum analysis, the highorder spectrum has good noise suppression and preserves the nonlinear phase information of the signal, which solves the problem of fast diagnosis of stator interturn short circuit fault. Through the theoretical analysis based on examples, it is shown that the bispectrum horizontal slice diagram at the "maximum amplitude" of the current signal with and without stator interturn short circuit fault is completely different from it at the "maximum amplitude" of the current signal without fault. With the increasing of the degree of the fault, there are more harmonic weights and more amplitude values. Therefore, it is proved that the method can effectively extract the nonlinear characteristics of the signal, complete the diagnosis of the stator interturn shortcircuit fault of the asynchronous motor and has good application prospect.