A novel parameter error driven robust estimation method is proposed for online estimation of centre of gravity (CG) height of a tour bus vehicle. Unlike the commonly used recursive least squares (RLS) method, a parameter error driven updating law is synthesised to ensure improved performance in terms of steady-state error and robust convergence. The stability of the proposed estimation method is illustrated using the Lyapunov approach. The proposed CG estimation method is based on directly measurable lateral acceleration and relative roll angle of the sprung mass, while the high order signals are eliminated using the filtering approach. The validity of the proposed estimation method is illustrated through implementations in TruckSim platform. Simulation results show that the proposed estimation method can yield accurate online estimation of the CG height, together with effective suspension roll stiffness and damping properties.