Abstract: In order to improve the prediction accuracy of ultra-short-term wind power,an ultra short term wind power prediction method based on adaptive lifting and prediction error correction is proposed. Firstly,the original wind power sequence is decomposed into multiple components using combining complete ensemble empirical mode decomposition with adaptive noise（CEEMDAN）,and reconstructed into new modes with refined composite multiscale entropy（RCMSE）to reduce the complexity of wind power sequence.Secondly,EESHHO is used to optimize the ELM weights and thresholds to improve the generalization of the model,and at the same time AdaBoost is introduced to improve the accuracy and stability of prediction model;Finally,the strategy to correct the prediction is proposed based on the historical training error value strategy,further improving the prediction accuracy. The results verify the effectiveness of the proposed method.