Abstract: With the high penetration rate of renewable energy in traditional power grid,the concept of virtual power plant（VPP）is proposed to integrate and utilize renewable energy. Energy collaborative optimization control method for energy storage system of virtual power plant is proposed based on model predictive control. Long-short term memory neural network is used to obtain the one day-ahead forecasting information,such as load,wind and photovoltaic within the jurisdiction of virtual power plant. With the minimum economic cost as the optimization goal,the optimal scheduling is solved by an improved particle swarm optimization algorithm under the framework of model predictive control. The effectiveness of the proposed scheme has been validated by simulation results.