Abstract: For the problem that traditional life prediction methods described product degradation stochastic process inadequately using monotonous regression function,the non-stationary time series types of different degradation amount distribution parameters estimation was analyzed based on product degradation amount distribution at each time.Different parameters estimation was modeled respectively using non-stationary time series analysis method.Hence,a realistic product life and reliability prediction method based on degradation amount distribution time series analysis was proposed. A degradation test of a certain electronic product was conducted,and product life and reliability were obtained by the suggested method. The results show that the suggested life prediction method is more reasonable than traditional life prediction method.