The randomness and volatility of wind power bring serious challenges to the safe and economical operation of power system. Reasonable wind power uncertainty model and unit commitment optimization method are the key to ensure the safety and economy of power system. In this paper, a two-stage stochastic optimization method for unit commitment of power system considering wind power is proposed. According to the nonparametric empirical distribution of wind power output history data, wind power dynamic scenarios are produced accorded with wind power randomness and volatility. Considering that some extreme boundary scenarios that are easily overlooked in the process of scenarios reduction will increase the risk of wind or load shedding, a two-stage optimization model of unit commitment with reduced scenarios and extreme boundary scenarios is proposed. At the same time, in order to solve the nonlinear mixed integer optimization problem of unit commitment, a hybrid genetic and crisscross optimization algorithm is proposed. The simulation results show that the proposed model and method are effective when solving the problem of unit commitment with wind power system.