WOFOST (world food study) model had been successfully used in daily business of agro meteorological monitoring and yield forecasting in European Union, and also been widely used in crop growth process simulation and yield estimation all over the world. In this study, with the help of the rice growth observed data, the meteorological data at the same time and the rice planting regional planning data in Heilongjiang Province, the crop parameters for WOFOST model were improved. Based on the localization and regionalization of the model, the rice yield in county and region scale in Heilongjiang Province was simulated. In province scale, the WOFOST simulated yield was good, and the relative error between estimated yield and statistical yield from 2006 to 2013 were respectively 2.71%, 8.47%, 6.41%, –15.96%, 3.95%, 0.02%, –7.06%, 0.88%, four of which beyond ±5%. But in county scale, the correlation between WOFOST simulated and statistical yield was poor, and not passing the test of significance. In order to improve the precision, the trend yield calculated by the statistical yield and the WOFOST simulated yield were both used to build a comprehensive rice yield simulation model by the multiple linear regression method year by year from 2006 to 2013. Then the rice yield both in county and province scale in Heilongjiang Province was calculated by using the comprehensive model. In county scale, the comprehensive simulated yield and the statistical yield in county scale passed significant test of 0.01, and the correlation coefficients were respectively 0.715, 0.728, 0.829, 0.810, 0.888, 0.919, 0.868, 0.798, the R2 were respectively 0.511, 0.529, 0.686, 0.656, 0.789, 0.844, 0.753, 0.636. In province scale, the relative error between the estimated yield and statistical yield during 2006–2013 were respectively –1.72%, 2.12%, 3.02%, –2.45%, 1.27%, –0.89%, –0.38%, 1.96%. The comprehensive model had a good effect on improving the defects of fluctuation in individual year with a relative higher accuracy than that of only using WOFOST model, and could satisfy the application of rice yield estimation in large region.