Abstract: In allusion to virtual power plants composed of distributed generators and energy storages, a two-stage robust optimal scheduling model, in which the risk was taken into account, was proposed to deal with the impact of distributed renewable energy uncertainty on the operation of virtual power plants. In the first stage of this model the uncertainty set of the robust model was optimized, meanwhile the risk of renewable energy curtailment and that of loss of load were evaluated. In the second stage, the worst operation scenario was searched within the optimized uncertainty set, and the operation cost of the virtual power plant under this scenario was optimized. Cooperatively solving the two stage models, an appropriate robust model uncertainty set could be determined and a dispatching scheme to cope with the worst operation scenario could be enacted to ensure both economy and security of virtual power plant operation. By means of the linearization of the two kinds of risks, the proposed model could be solved by the state of the art column-and-constraint generation algorithm. Finally, the effectiveness of the proposed model is verified by case study.