Abstract: Based on the analysis of research status and in order to solve the problem of the low precision of irrigation water consumption forecasting, chaotic forecasting model was put forward. The evolvement regulation was resumed in high dimension space with state space reconstruction of irrigation water consumption serials, and the chaotic evolvement characteristic was distinguished. The chaotic forecasting model was established with the maximal Lyapunov index. The model was applied to forecast irrigation water consumption of Jushui River basin in Hubei Province. The comparison of forecasting results which calculated by using chaotic forecasting model with witch calculated by using other methods demonstrated that the proposed model had better forecasting effect, and it could be regard as an effective forecasting method for agricultural irrigation water consumption forecasting.