Stabilization of a plant in case of uncertainty parameters and unmodeled dynamics are the main problems considered in this paper. A robust control of motion of a rigid shaft that is supported by magnetic bearings was used as an example. The dynamics of the active magnetic suspension system is characterized by instability and uncertainty. The uncertainty is modeled as an additive and multiplicative. Robust controller H∞ was designed for the defined plant with the uncertainty models. The robust controller assures high quality of control despite the uncertainty models. Robust control of vibrations of a rigid rotor is confirmed by experimental studies. A digital signal processor is used to execute the control algorithm in real time.