With the increase of wind power penetration, the correlation of wind speed among wind farms should be considered while considering the randomness and intermittency of wind power. In this paper, Rosenblatt transformation and Copula function are combined to deal with the complex nonlinear relationship among wind farms, and a probabilistic load flow cumulant method based on Copula function and Rosenblatt transformation is proposed. The method can capture the nonlinear correlation among input variables accurately and has the advantage of high efficiency of cumulant method. The IEEE33-bus system including wind farms is used for testing the proposed method. The results verify the validity, accuracy and practicability of the proposed method, and the calculation results of the proposed method are closer to the actual operating condition than that of the method which considers linear correlation. This work is supported by National Natural Science Foundation of China (No.51467012).