In terms of an Artificial Neural Netework(ANN),a forecast and diagnostic analysis system of the farmland soil moisture is established.A mixed model and a model involving the extension of time series of the artifivial neural network may be given in the system,respectively,based on soil humidity change and physical factors influencing its change as well.This system may be used to estimate depth of infiltrating water into soil and the change of soil moisture of different soil layers.Application results show that the system operated in interactive mode with Chinese character menu is convenient and superior on prediction and diagnostic analysis for farmland soil moisture due to ANN marked by quasidynamical characterstic,adjusting parameter and structure.The sysytem is suitable to relieve the calamity of the agrometerological drought and to manage water resources.