An objective correction method is established by making use of ensemble forecast of 10 m wind speed from 2013 to 2015 and wind speed observation from 88 representative stations. The accuracy and stability of the mode are equal or greater than the average values. Thus fully considering the prediction probability of historical data and collecting the distribution of each member of ensemble forecasts could improve the correction effect. Small error and TS score are significantly improved by correction. Specifically, TS score of 8-9 magnitude of wind speed forecast within 72-120 h lead time increases from 0.04 to 0.44, which can effectively promote the forecasting ability of medium and longterm timeliness and high volume of wind speed. The corrected forecasts could bring good effectiveness for cold air and typhoon weather in the offshore areas of China.