Abstract: In terms of leakage circuit breaker, it is a key terminal protector in power system. To avoid its excessive usage as well as ensure the safety of life and property of power users, it needs to carry out the research on its life prediction method. A set of life prediction method based on statistical test is proposed. This method makes the combination of statistical theory with Arrhenius model, which solves the applicability of Arrhenius model through normality test and correlation test. At the same time, the aging mechanism of accelerated aging test is solved through variance homogeneity test. In addition, the quality problem of the fitting model is solved and the accuracy of the prediction model is further improved by heteroscedasticity test, residual normality test and residual independence test. The proposed method is applied to make the life prediction of electronic leakage circuit breaker. The prediction results are close to the reference life provided by the manufacturer, which indicates that the proposed method can solve the problems existing in the life prediction process in an effective way. At the same time, a certain reference value of the proposed method can be provided for the life prediction of other similar products.