Abstract: In Shanghai, the electricity load density is relatively high and the distributed power resources are abundant. However, the distributed power access lacks unified planning. This article analyzes the coordinated planning of Shanghai power supply based on power source, grid and load, together with the resources required. A collaborative planning model of power source, grid and load is established which is solved by two-level cross feedback particle swarm optimization simulated annealing algorithm. This paper evaluates the power distribution network planning results before and after considering the collaborative planning of power source, grid and load. It shows that the planning method mentioned in this paper benefits in several aspects as better operation economy, characteristics and efficiency of distributed power utilization.