The invention discloses a neural network based multidimensional-parameter intelligent crop-irrigation control method. According to the method, parameter information of current field to be irrigated, including rainfall, soil moisture, wind velocity, humidity and temperature, illumination intensity, flow rate and the like, are acquired, a crop water requirement model taking a crop water requirementsignal as a response message is built on the basis of the neural network, and is used for computing the multidimensional environmental parameters, the crop water requirement of the current field is then predicted, and a controller makes a comprehensive judgement result according to the water requirement, the rainfall and the soil moisture, and controls an electromagnetic valve according to the judgement result; thus, accurate irrigation to crop is completed without manual operation in the whole process. The control method has the advantages of having high work efficiency and high precision, increasing utilization rate of water resource, crop yield and agricultural information degree, reducing production cost of crop and labor cost, having a positive promotion effect to improve crop qualityand increase crop yield, and having obvious social benefit.