The invention provides a method for predicting generated energy of a photovoltaic power station. By means of environment monitoring data and operating data in an area, where the photovoltaic power station is, a regression model is fitted by using a multiple regression technology, such that prediction sample data are obtained; the prediction sample data are continuously compared with historical data and a historical prediction result and improved; and thus, relatively-low error prediction of the generated energy is realized. Prediction of the generated energy of the photovoltaic power station is completed; the generated energy can be predicted in advance; the operation stability of a power grid is improved; the data accuracy can also be checked; and the data quality is improved.