Abstract: A short-term wind speed combination prediction model based on complementary ensemble empirical mode decomposition and Stacking fusion is proposed to address the issue of low accuracy in short-term wind speed prediction in wind farms. Firstly, in order to highlight the local characteristics of short-term wind speed and reduce modeling difficulty, the short-term wind speed is decomposed into several stable components using complementary ensemble empirical mode decomposition algorithm. Then, information entropy and approximate entropy are used to determine the complexity of each component. The high complexity component selects the least squares support vector machine, and the low complexity component selects the stochastic configuration networks as the corresponding prediction model. By using the Stacking algorithm to fuse the predicted values of each model, the prediction accuracy is improved. Finally, using a group of actual short-term wind speed data set as the research object, the proposed prediction model is applied to its prediction. The comparison results indicate that the proposed prediction model improves the accuracy of short-term wind speed prediction.