The structure of this paper is as follows: the second section mainly introduces the least square support vector machine algorithm and error correction principle; The third section introduces the modeling data and evaluation index of two wind farms; the fourth section describes the power of the prediction results and error analysis of the simulation of two wind farm power data at home and abroad and the fifth section summarizes the full text and makes a prospect.

Because the traditional power prediction method cannot met the prediction accuracy, and it had been found that the error of adjacent time is similar to the error level, so the error correction was introduced into the LS-SVM model in this paper to establish the least square support vector machine power prediction model based on error correction and the power data of two wind farms in Spain and China were used for simulation.