Abstract: With the development of multi-region interconnected power system, the wind power, photovoltaic (PV) and other new energy power generation are connected to the grid on a large scale, and it is an inevitable trend and urgent need to perform the joint prediction and coordinated scheduling of wind power-photovoltaic power. Starting from the analysis of the spatio-temporal correlation of wind power-photovoltaic (PV) power generation, this paper analyzes and discusses the spatio-temporal coupling of wind power-PV power, modelling of wind power-PV joint forecast, optimization of the parameters of wind power-PV joint forecast model, and pre-dispatch of the power system considering the wind-PV joint forecast. Firstly, the mechanism of wind-PV power interaction in time-space is revealed, a multi-time scale wind-PV power time complementarity analysis method is proposed, a quantitative model for the spatial correlation of wind-PV power generation is established, and a spatial and temporal coupling model for wind-PV power generation is constructed based on a multistage graph convolutional neural network. Based on the results above, a joint prediction method of wind-PV power by integrating heterogeneous graph neural networks is proposed, a joint forecasting model for wind-PV power optimization model of the parameters of the joint prediction model is constructed, and a new energy active power prediction error vector evaluation system is constructed, which provide decision-making supports for the coordinated scheduling and control of the wind-optical joint power generation system. On the basis of wind-optical joint power generation prediction, the wind power and photovoltaic power generation time complementary and spatial mutual aid ideas are used to explore the wind-optical joint power system pre-scheduling strategies and methods, and the wind-optical coordinated scheduling strategies at different time scales are dissected and analyzed, and a hierarchical scheduling framework for the combined wind-photovoltaic power generation system with spatial progression of power grid-region-cluster-field station is established. Finally, the prospects in future researches on wind-PV joint prediction and pre-dispatch are put forward.