It is significant to constrain the ramp rate of wind farms when large-scale wind generation is integrated in the grid. We proposed a novel method based on probability prediction of wind generation using energy storage system（ESS） for ramp rate control in the wind farm. The Gaussian process regression（GPR） model was used to predict the next step generation of the wind farm. When the predicted ramp rate exceeded the limited value

energy storage was used to compensate the variation. The historical data of a wind farm were used for case study. The calculation results show that the accuracy of the one-step GPR prediction can satisfy the requirements of ramp rate control. Considering the restriction of the ESS capacity

the amount of ramp events can decrease by 50% or more. The work of the paper can provides an effective way for wind farm ramp rate control using energy storage which can be used in the practice.