Abstract: The operation mechanism of the virtual power plant can provide the technical support for the integration of renewable energy and distributed power on the demand side. In this paper, a two-stage robust optimization model is built for virtual power plant (VPP) which includes wind power units, thermal power units, hydro units and flexible loads. Based on wind power completely consumed, the operation plan of each unit is optimized to alleviate the uncertainty and randomness of wind power, and guarantee the optimal operation efficiency of the virtual power plant. To solve the two-stage problem with min-max-min structure, it is decomposed into a main mixed integer problem and a linear optimization sub-problem which considers the uncertainty of wind power output. Then, the linear optimization strong duality theory and the column constraint generation algorithm(C&CG) are introduced to solve the main problem and the sub-problem by iteration. In the iteration process, the dual model of the sub-problem is linearized by the Big-M method. Finally, the effectiveness of the proposed method is verified by experiment results.