Abstract: Using the mean monthly air temperature from 1961 to 2000, the corn heat indexes in 3 provinces and whole Northeast China are calculated. Based on the statistic analysis between heat index and atmospheric circulation data, 6 models of forecasting availably the corn heat index are set up. Each model can be used to forecast heat status in the growth and development period of corn in each province and whole Northeast China. Test shows that the accuracy of all models is higher and the stability is preferable. The mean relative error of 6 models is lower than 7%. These models show the better forecasting ability. In which, the least relative error is in Liaoning Province, the mean value is lower than 2%, and the forecast effect is best. The greatest relative error is in Heilongjiang Province, the mean value is from 4.5% to 7%.