| The movement of soil moisture is influenced by multiple factors, such as weather elements, soil characteristics and crop growth, etc. Based on the observation data of soil water content and meteorological data during 1993 to 2001, a back-propagation (BP) neural network model for soil water content forecast is established. The application results show that the predicted values are coincident well with the observed values, and it is feasible to apply the model in predicting soil moisture condition and setting up irrigation plans. |