A model of wind power curve is developed from industrial data.
Wind power and speed data is clustered.
Support vector machine algorithm is applied to construct the model.
The model captures dynamic behavior of the wind turbine.
Computational results demonstrate effectiveness of the proposed approach.