In view of the flaws of slow dynamic response of photovoltaic grid-connected inverter system and poor grid-connected current tracking effect of traditional PID control, a control method of BP neural network combined with PID has thus been proposed. With single-phase LCL grid-connected inverter being the research object, an analysis has been made of the circuit structure and BP neural network model of single-phase LCL grid-connected inverter. By adopting the back-propagation BP learning algorithm with a momentum update, the convergence of error performance function is accelerated, followed by a real-time quick output of appropriate PID parameters, thus improving the response speed of the system. Finally, the Matlab model is constructed for simulation. The simulation results show that compared with the traditional PID controller, the control strategy of BP combined with PID is characterized with a better performance in completing the grid-connected current tracking, with a faster speed as well as a smaller steady-state error, thus verifying the effectiveness of the method.