Abstract: A short-term photovoltaic power prediction method based on BP neural network with similar days and momentum optimization is proposed. The similar days are selected by using the irradiance which is strongly related to the output power as a similar variable.The BP neural network prediction model is established by using the momentum method and the historical data and meteorological information of similar days as training samples. The actual operation data of a photovoltaic power station in Xinjiang are used for verification and analysis. The results show that the method can achieve accurate prediction accuracy in sunny and non sunny weather,and verifies the accuracy and effectiveness of the proposed model and algorithm.