Abstract: Based on the laser wind radar data,and aiming at the nonlinear characteristics of wind speed,the Sparrow Search Algorithm（SSA） is proposed in this paper to optimize the Extreme Learning Machine（ELM） for wind speed forecast.Based on the forecast model established,the pre-pitch is performed according to the predicted wind speed,and the moment load of the wind turbine blade root is analyzed. The simulation is carried out based on the laser radar data of the wind speed measurement of a wind farm in Xinjiang and comparison is made with other prediction models. The results show that the extreme learning machine optimized by Sparrow can accurately predict the wind speed,and significantly improve the prediction speed of the extreme learning machine and the dynamic performance under different wind speed conditions;after the pre-pitch,the blade root moment load of the wind turbine is greatly reduced,which improves the service life and operational safety of the blades.