ABSTRACT The maglev rotor installed on a moving base is subjected to both the unbalanced force of the rotor itself and the base excitation caused by external interferences. The interaction between the maglev rotor and the moving base gives rise to a combination of challenges. Specifically, the base excitation can propagate to the rotor, inducing vibration, while the unbalanced vibration can reciprocally transmit to the base, thereby augmenting the complexity of the base vibration. This paper introduces a feedforward decoupling control approach to address the base excitation vibration of the maglev rotor, utilising the base acceleration as the reference signal. The foundation of the feedforward controller model for countering base excitation vibration is constructed through the utilisation of inverse system principles and system identification techniques. The resultant decoupling model is engineered to facilitate decentralised control of the rotor across four radial directions. Simultaneously, to mitigate the impact of rotor unbalance on the base acceleration reference signal used by the feedforward controller, the least mean square (LMS) adaptive filtering algorithm is employed. This strategy facilitates unconstrained rotation of the rotor around its inertial axis, thereby reducing the unbalanced inertial forces and enabling isolated control of both base excitation vibration and rotor unbalance-induced vibration in the maglev system. An experimental device is constructed to validate the effectiveness of the proposed control methodology. The experimental results show that the maximum vibration amplitude of the maglev rotor decreases by 47% ~71% after the control method is applied.