Solar irradiance is one of the most crucial factors affecting the photovoltaic (PV) power output. Accurate solar irradiance forecasting can help predict future PV power generation for grid managers, thus to mitigate the adverse effect of random PV power on the power grid. However, research on real-time surface irradiance distribution in a certain region still not be presented in previous studies, which can bring further convenience to solar power forecasting and power grid dispatching work. In this paper, a surface irradiance distribution mapping model for PV power station based on ground-based sky images is proposed. First, we process raw sky image with distortion rectification and background message dislodged. Then after cloud motion displacement calculation by hybrid Fourier phase correlation (HFPC) theory, we extract the gray level and position information of pixels in sky images to explore the mapping relationship between sky image and solar irradiance. Subsequently, an adaptive sky image-irradiance mapping model is built, trained, and updated according to real-time sky images and solar irradiance. Simulation results show that the proposed method in relation to forecast can achieve real-time surface irradiance distribution mapping at quite high accuracy.