The objective of this study is to develop rice response models to account for the impact of weather on rice yield and to project the rice yield level. In order to consider spatial autocorrelation effects, spatial econometrics approaches are used. The empirical results showed that rice yield level are strongly influenced by weather variables. The forecasting power of the rice yield models was improved by adding the spatial autocorrelation effect. Out of sample forecasting tests confirmed the spatial response models are more accurate than using general multiple regression models.