Abstract:Based on the data of China’s grain output from 1978 to 2018, the combinatorial prediction model based on IOWA operator is proposed to predict China’s grain output in the next five years. Firstly, time series ARIMA model, Holt-winters addition model and multiple linear regression model are set up, and three single prediction models are used to fit China’s grain output from 1978 to 2018. Then Based on IOWA operator, the combinatorial prediction model by minimizing error sum of squares as the criterion is established. Finally, the combinatorial prediction model based on IOWA operator is used to forecast China’s grain output in the next five years. The conclusion shows that the combinatorial prediction model based on IOWA operator has higher prediction effect, can overall use the effective information provided by single prediction model and can more accurately predict China’s grain output. In the next five years, China’s grain output will continue to increase, and the average annual grain growth rate is about 1.3 percent.