Li-ion batteries are at the core of new energy vehicle powertrains. The model-based battery management system (BMS) is the key to ensure the full play of battery performance. However,the existing BMS mainly adopts the equivalent circuit model (ECM),without considering the impact of discharge rate on available capacity. Therefore,the model will have significant terminal voltage simulation errors in different discharge rates and low state of charge (SOC) regions,which affects the accuracy of BMS. BMS cannot accurately estimate the battery discharge cut-off condition,and the estimation error of residual discharge capacity (RDC) is large,which may lead to serious consequences such as battery voltage sag and vehicle breakdown. To solve the above problems,based on the extended equivalent circuit model (EECM) considering the internal diffusion mechanism,the incremental capacity (IC) curves of discharge voltage at different rates are compared and analyzed. Nernst equation is used to construct capacity open circuit voltage curves at different discharge rates,and an improved EECM is proposed. The simulation errors of the improved EECM for terminal voltage under different current rates and dynamic conditions are smaller than traditional ECM and EECM. Improved EECM can improve the accuracy of RDC estimation and have the potential to be applied in practical BMS.