In order to reduce the impact of scaled distributed power access on the original grid and improve the multiple utilization efficiency of the source grid.In this paper,by analyzing the multi-stage energy interaction scene with distributed power supply,the multi space scale of main network,distribution network,feeder and platform area,and short-term,ultra short term,quasi real time multi time scale interaction mechanism are formulated,a multi-source interactive coordinated optimal control model for distributed power supply including large wind power plants,gas turbines and centralized large capacity energy storage is established.The objective function of the model is the minimum active power fluctuation,the closest network loss and the largest economic benefit.The model sets constraints such as node power flow,wind power,and so on.Taking Chongming Island in Shanghai as an example,the multi intelligence coordinated optimization control application scene of source network is set up to simulate and verify the short-term,super-short-term and quasi-real-time optimization effects of multiple spatial scales,and to provide optimized control basis for the access of large-scale distributed power supply.