Battery state of charge (SOC) is a critical parameter for the control of propulsion systems in plug-in hybrid electric vehicles (PHEV) and electric vehicles (EV). As SOC is not measureable during vehicle operation, an onboard adaptive algorithm is developed in this paper. The algorithm estimates six electrical parameters for Li-ion batteries and provides a reliable SOC based on one of the estimated battery parameters, i.e. open circuit voltage (OCV). Simulation and vehicle validation results show good robustness and adaptation of the algorithm with high computational efficiency and low implementation cost.