Regional crop yield prediction is a vital component of national food security assessment. Data assimilation method which combines crop growth model and remotely sensed data has been proven the most potential method in regional crop production estimation. This paper takes Hengshui district as study area, WOFOST as crop model, MODIS-LAI as observation data to test and verify the efficiency of EnKF assimilation method. The results show that the precision of crop yield estimation are obviously improved with EnKF assimilation, in the WOFOST potential level the R2 improved from 0.10 to 0.38 and RMSE was reduced from 2480 kg/ha to 880kg/ha. Our study indicates that EnKF assimilation method has great potential in regional crop production forecasting.