The integrity of wind data is of great significance for the accurate prediction of wind power and the utilization of wind energy. Based on the correlation analysis of the output power of the wind farm, this paper proposes a joint probability distribution model of different wind turbine output power by using Copula theory. By using the measured wind power data, the probability distribution of wind power output is estimated through the non parametric and density estimation method. Kendall rank correlation coefficient is used as the correlation measurement. The correlation between different wind turbines is used to complete missing data. Simulation experiments illustrate the accuracy and average relative error of completed data obtain good effect, and the quality of the completed data is improved effectively.