The existence of distributed power supplies, energy storage devices, flexible loads, and other multi-agents with uncertain output in smart parks poses great challenges to traditional optimal regulation. This article proposes an optimization strategy based on the accelerated alternating direction method of multipliers (A2DM2) to solve the distributed collaborative control problem of multi-agent smart parks. Firstly, the smart park system is divided rationally, and the minimum cost is established as the goal of the distributed control model. Then, according to the multi-agent output constraint, the A2DM2 is used to optimize the solution. Finally, it is verified through simulation.