Abstract: Under the background of high proportion of renewable energy access, the randomness of wind and solar output and load fluctuation bring challenges to the economic dispatch of virtual power plants. Therefore, an optimization strategy based on the collaboration of proximal strategy optimization (PPO) and CPLEX solver is proposed. The strategy uses the PPO algorithm to generate wind-solar load scenarios and screen out typical scenario sets, and uses CPLEX to construct a mixed integer linear programming model with the goal of minimizing daily operating costs and strictly satisfying power balance and energy storage state of charge (SOC) constraints. The results show that the strategy proposed in this paper can significantly reduce the operating cost, verify its feasibility, economy and practicability in dispatching, and provide an efficient solution for high-proportion renewable energy grid.