The support vector machine (SVM) is a novel type of learning machine based on statistical learning theory

that can map a nonlinear function successfully. As a battery is a nonlinear system, it is difficult to establish

the relationship between the load voltage and the current under different temperatures and state of

charge (SOC). The SVM is used to model the battery nonlinear dynamics in this paper. Tests are performed

on an 80Ah Ni/MH battery pack with the Federal Urban Driving Schedule (FUDS) cycle to set up the

SVM model. Compared with the Nernst and Shepherd combined model, the SVM model can simulate

the battery dynamics better with small amounts of experimental data. The maximum relative error is 3.61 .