With the increasing proportion of distributed renewable energy access to the power grid, the intermittent and stochastic output of power system has brought a series of negative effects on the safety, stability and economic operation, and the Virtual Power Plant (VPP) provides a new way for the reliable interconnection of distributed renewable energy sources. Aiming at the uncertainty and prediction error of the output of the stroke generator in the VPP, this paper considers that several wind power forecasting service providers provide a number of air power output prediction scenarios based on different forecasting methods, and the virtual plant sets the optimal scheduling plan for each scene with the goal of minimizing the expected operating cost. In response to random fluctuations in wind conditions, the current time slot of a VPP considers the planned adjustment costs under all wind power output scenarios in the next period. Aiming at the minimum total cost of expected operation, the flexible optimal scheduling model of VPP based on multi-wind energy prediction results is constructed, and the rolling scheduling strategy of the output plan of intraday VPP is designed. Through numerical simulation, the running cost of VPP under different prediction scenarios is compared and analyzed, and the feasibility and effectiveness of the proposed model are verified. This work is supported by National Natural Science Foundation of China (No. 71420107027), Science and Technology Projects of Hunan Province (No. 2017CT5015 and No. 2017WK2053), Postgraduate Research and Innovation Project of Hunan Province (No. CX2017B465), and Strategic Emerging Industries Project in Hunan Province (No. 2018GK4002).