Battery management system (BMS) is an important part for battery energy storage system to guarantee the safety operation. It is of great significance for the operation and maintenance of the energy storage power station to test the system before the BMS is put into operation. However, there is no test specification and standard for the battery energy storage system BMS in the field of state of charge (SOC) estimation at present. Therefore, this paper establishes a test platform for BMS of battery energy storage system, Thevenin equivalent circuit model based on the information of external characteristics of battery is utilized, and the method of extrapolation of battery multiple discharge curve is used to obtain the open circuit voltage curve of battery. Besides, the dual extended Kalman filter (DEKF) algorithm is proposed to realize the accurate estimation of SOC. By comparing with the EKF method, DEKF method has advantages in convergence speed and SOC estimation accuracy. Under the typical federal urban driving schedule (FUDS) and dynamic stress test (DST) conditions, the SOC error estimated by DEKF method and EKF method is less than 1%. The battery terminal voltage error is within ±10 mV and ±20 mV respectively, and the average absolute error is within ±20 mV respectively 2.7 mV and 3.8 mV.