Abstract: A combination forecasting model was applied to forecast the grain production in Fujian Province and to improve the precision of the prediction.Five evaluation indicators were used to comprehensively compare the different forecasting methods.The weight of each model in the combination forecasting model was objectively calculated with entropy weight theory.Three production forecasting models were combined to establish a combination forecasting model for the grain yield by proper weights.The forecasting results of all the methods were weighted and summed to gain the final forecasting results.The combination forecasting model was used to simulate the crop yield from 1949 to 2004 in Fujian Province.The result showed that the combination forecasting method would increase forecasting accuracy and reduce risks at the same time.The forecast precision was obviously improved.

Key words: Grain production, Grain production, Combination forecast, Entropy weight, Fujian Province