Abstract In this paper, a fuzzy torque vectoring differential controller is proposed to improve the lateral stability for the two-motor-wheel drive electric buses with low sensor cost and computational cost. In order to know the unmeasured sideslip angle, unscented Kalman filter algorithm is used to accurately estimate the sideslip angle based on a nonlinear electric bus model and the measured yaw rate and lateral acceleration signals. Given the advantages of high computational efficiency and human’s heuristic knowledge, fuzzy rules are newly designed to control the torque vectoring differentials, and thereby to provide an extra yaw moment on the electric bus body to compensate the driver steering aiming at stability enhancement. To validate the effectiveness and efficiency of the proposed fuzzy torque vectoring differential controller, hardware-in-the-loop tests were carried out to evaluate the controller performance in real-time under different driving maneuvers and road conditions. The test results indicate that the fuzzy torque vectoring differential controller can significantly improve the handling and path-following accuracy for the two-motor-wheel drive electric buses.