Abstract: Solar photovoltaic technology is becoming one of the hot issues in the field of renewable energy generation. In future, China’s large-scale grid-connected photovoltaic power generation system will be continuously in rapid development. But, so far, the exploring of photovoltaic power generation forecasting is still weak, and there are few methods available to meet the practical needs of photovoltaic power generation prediction in China. Photovoltaic power generation prediction means to accurately predict solar irradiances at first, and then in combination with the analysis of the historic power generation data of photovoltaic power station, to further forecast photovoltaic power. In this paper, we briefly introduce and classify several types of photovoltaic power generation forecasting models, such as the simulation-prediction method based on global solar radiation prediction and photovoltaic simulator, the physical prediction method based on global solar radiation prediction and photoelectric conversion efficiency model, the statistic-dynamic method based on the meteorological data and photoelectric power generation data processing and numerical weather prediction. Meanwhile, we also simply introduce photovoltaic power generation forecasting platform’s construction abroad.