Abstract: Continuously improving the prediction accuracy of the characteristic quantities of wind power ramp events is of great significance to the safe and stable operation of the power system. Therefore, this paper proposes a rolling correction model of ramp for wind power that combines the characteristic quantities of ramp events with meteorological data of numerical weather prediction（NWP）. First, based on the parameter and resolution adaptive（PRAA） algorithm, all the characteristic quantities of ramp events in the historical database and the prediction database are obtained, and the forecasting error vector matrix of ramp characteristic quantities is established. Then, the linear and non-linear correlation relationships of the error vector matrix and the meteorological data in NWP are analyzed and effective meteorological indicators that affect the prediction error of the ramp characteristic quantities are identified. Finally, effective meteorological indicators with the similar characteristic in future and historical databases are matched based on dynamic time warping（DTW）. The correctional reference set of predicted errors of future ramp events is obtained and rolling corrected. The calculation example shows that the correction model can effectively reduce the ramp amplitude error and improve the prediction accuracy of ramp events.