Abstract: Short-term and ultra-short-term wind power prediction（WPP） in regions is an effective method to enhance the robustness of the power grid. The state of the art of short-term and ultra-short-term WPP techniques was summarized in the paper. The classification of WPP techniques was discussed, in terms of prediction method of single wind farm and regions. The overall framework of the regional WPP system was discussed in four aspects, including the forecasting flow chart, data sources, data flows and the physical levels. Generalization physical layers of wind power forecasting was presented in five levels, data layer, mapping layer, feature layer, methodology layer, and feedback layer, based on which the improvement methods for WPP were discussed. The proposed improvement methods for the five physical levels of WPP will contribute to both short-term and ultra-short-term WPP in regions.