Based on BP Neural Network Forecasting Model of Soil Water Storage

[Objective] The study aimed to provide the scientific basis for realizing the real-time irrigation of crops.[Method]The BP neural network forecasting model was established by the measured meteorological data,2008-2009 winter wheat test data of water-saving irrigation Huantai County Experiment Station.Simulated training was conducted by applying neural network toolbox Mat lab and rainlm,at the same time soil water storage was predicted.[Result]The results showed that the BP neural network forecasting model had a great generalization ability,the model had a higher forecast accuracy and better stability in stage of jointing,blooming and grouting of winter wheat.[Conclusion]The simulation value had a high accuracy in stage of a big soil water storage of winter wheat.