Abstract: The utilization rate of fertilizer in China ranks first in the world all the year round, and the utilization rate of agricultural water is low. The method of relying on personal experience not only wastes fertilizer and water resources, but also pollutes the local ecological environment. Because of pipeline transportation and other reasons, water-saving irrigation and fertilization system has the characteristics of time-varying, nonlinear and time-delay, so it is difficult for ordinary controllers to accurately control the flow of water-saving irrigation and fertilization system. Aiming at the above problems, this paper studies and designs a controller based on BP neural network PID, in order to realize the precise control of liquid fertilizer flow in water-saving irrigation and fertilization system, and compares it with the traditional PID controller, and obtains the step response curve by simulation analysis with Matlab software. The results show that the controller based on BP neural network PID has excellent control effect, which can meet the actual requirements of water-saving irrigation and fertilization system to achieve accurate control.