This study presents a new equivalent lithium-ion (Li-ion) battery model for online energy management system. It has an equilibrium potential E and an equivalent internal resistance Rint. The equilibrium potential E is expressed as a function of state-of-charge (SOC), current and temperature. The equivalent internal resistance Rint includes R1 and R2. R1 is defined as the resistance, which can be formulated by the discharging current and temperature. R2 is defined as the resistance which is because of the change of temperature. The adaptive extended Kalman filter is employed to implement the online energy management system based on the proposed Li-ion battery model. The SOC is considered as the state variable for the charging or discharging process of the Li-ion battery. The covariance parameters of the processing noise and observation errors are updated adaptively. The SOC of the Li-ion battery can be predicted by the online measured voltage and current in the online energy management system. The effectiveness and robustness of the proposed Li-ion battery model is validated. Experimental results show that the estimated SOC is accurate for various operating conditions. A comparison between the proposed method and other SOC estimation methods is also shown in the experimental results and analysis section.