The energy crisis is getting more and more serious around the world. In recent years, renewable energy has developed rapidly as one of the ways to relieve the energy crisis. However, renewable energy has the characteristics of uncertainty. Therefore, high renewable energy penetrated power system will bring new challenge to the optimal dispatching for the power system. In the case of rapid development of renewable energy power generation and nuclear power station, the peak and valley difference of the power system becomes larger, which increase the pressure on the power system peak regulation. In order to increase the peak regulation capacity, a model for flexible participation in peak regulation is proposed in this paper. Meanwhile, a two-stage stochastic optimization for power system nuclear power plants participating in peak regulation with the consideration of the stochastic renewable energy output is given. The first stage of optimization determines the status of the power plants. In the second stage of optimization minimize the expected operation cost of power system based on the decision made in first stage. The model is tested by IEEE 24 bus system and a real power system. The results verify the feasibility and economy of the nuclear plants participating in peak regulation.