State of charge (SOC) forecasting model, which are based on Back-Propagation neural network algorithm, was given for Solar PV System. Matlab programs was wrote to check the performance of the three layers BP network model, and corresponding experiments was done，Based on that, this method utilized TMS320F28335 as MCU of the hardware circuit to gather real-time data, and utilized the predicting value of SOC and control circuits to realize the intelligent monitoring of Valve-Regulated Lead Acid Battery pack, and ensured the economic, efficient, safe and reliable operation of the system. The monitoring system was Provided with SOC prediction, charge and discharge current parameters real-time monitoring, data transmission and status display, etc.