Aiming at the difficulty in obtaining sufficient failure data for the lifetime assessment of traditional high-reliability electronic products, a reliability evaluation and lifetime prediction method for photovoltaic inverters based on stochastic process theory is proposed in this paper. Firstly, the photovoltaic (PV) inverter performance degradation data is obtained by accelerated degradation experiment, as well as the output power is selected as its performance parameter. Secondly, the performance degradation model is established by the typical wiener process theory, and the maximum likelihood and Bayes method is used to estimate the degradation model parameters. Finally, lifetime estimation and remaining useful life prediction are realized according to the PV inverter failure threshold. The experimental results verify the validity and accuracy of the proposed method.