Abstract: This paper put fooward one kind of EMD-ELD wind power prediction methods based on hierarchical clustering method for solving the problem that the wind power prediction is not accurate. Make the datas like weather information into cluster by aggregation algorithm of hierarchical clustering method. Use the EMD method tools decompose the power sequence of each group. Then the datas are relatively stable. At last use the ELM model to calculate and reorganize the components. The characteristics of weather datas isn't obvious, so the prediction through aggregation algorithm is more accurate. The example simulation shows that this model is more accurate than the traditional prediction model.