Abstract: Large-scale wind power grid integration will bring great challenges to the normal operation of the power system. Accurate wind power prediction can improve the level of wind power consumption and ensure the normal operation of the power system. To address an insufficient prediction accuracy in sudden changes in wind power,a short-term wind power integrated prediction method based on power ramp feature identification is proposed in this paper. First,based on the historical data of wind power,the power sequence is divided into uphill climb,downhill climb and non-climb by using the climbing definition. Second,considering the characteristics of different climbing sections,the prediction method adapted to its characteristics is selected: SSA-BiLSTM model,CNN-BiLSTM model and LSTM model are used to predict the upper,lower and non-climbing stages, respectively. Finally, the prediction results of each segment are linearly superimposed to obtain the prediction results of wind power in the entire period. Taking the power sequence provided by a wind farm in Belgium as an example for verification,the results show that,compared with the traditional prediction model, the prediction method proposed can improve the accuracy and precision of wind power prediction effectively.