Wind speed dependences on different areas in a wind farm have influences on security and economic operation in power system. In order to simulate the correlation of wind speed series between different positions, this paper applies Copula function and rank correlation matrix methods to measure the coherence of wind speed in a wind farm. The correlated wind sample space is established. According to active power output characteristics of wind turbines, the polymerization model in a wind farm can be achieved. Monte Carlo optimal power flow is applied to IEEE-30 and IEEE-300 bus systems based on the principle of energy saving dispatching. The study shows that the accuracy of outputs is improved, thus reducing the fluctuation ranges in unit generating costs and power flow in branches while considering wind speed polymerization. This approach provides a new method to improve the effectiveness of energy saving dispatching and system operation arrangement. Results have been tested to be effective.