The basic task of a battery management system (BMS) is the optimal utilization of the stored energy and minimization of degradation effects. It is critical for a BMS that the state-of-charge (SoC) be accurately determined. Open-circuit voltage (OCV) is directly related to the state-of-charge of the battery, accurate estimation of the OCV leads to an accurate estimate of the SoC. In this paper we describe a statistical method to predict the open-circuit voltage on the basis of voltage curves obtained by charging batteries with different currents. We employ a dimension reduction method (Karhunen-Loeve expansion) and linear regression. Results of our modelling approach are independently validated in a specially designed expertment. (c) 2005 Elsevier B.V. All rights reserved.