An adaptive back-stepping control (ABC) is originally applied to a linearized model of an active magnetic bearing (AMB) system. Our control objective is to regulate the deviation of the magnetic bearing from its equilibrium position in the presences of an external disturbance and system uncertainties. The ABC consists of a recursive Lyapunov controller and adaptive laws. It is based on full-state feedback, for which all three states in the linearized AMB model (velocity, position, and current) are estimated to construct the control law. It is proved that the adaptive back-stepping controlled AMB system is asymptotically stable around the system's equilibrium points. Simulation results verify the effectiveness and robustness of the ABC against external disturbances and system uncertainties.