Abstract: To address cumulative errors in state of charge estimation and poor system reliability of vanadium redox flow batteries, a battery management system was designed based on the Simulink model-driven development method. A distributed hardware topology was constructed to integrate data acquisition and fluid pump control modules. Software algorithms were modeled via Simulink. A closed-loop SOC estimation algorithm fusing bypass open circuit voltage and ampere-hour integration was proposed to dynamically calibrate errors. Additionally, multi-stage hysteretic thermal management was designed. System tests indicated stable concurrent communication and qualified CPU real-time load. The proposed algorithm effectively suppresses zero-drift under complex conditions and significantly improves SOC accuracy. This system improves software development efficiency and ensures the safe and precise operation of VRFBs.