This paper is concerned with the effects of magnetorheological (MR) fluid damper model uncertainties on system stability, when local feedback linearization schemes are utilized. The objectives of the research are to characterize such effects, and develop a robust control law for given uncertainty bounds. First, deterministic functions are used to represent the possible deviations between the actual MR damper force and the force predicted by the model. Limit cycle behavior is predicted using the Describing Function Method and parameters influencing the limit cycle characteristics are identified. An effort is then made to derive the worst-case model deviation, given the bounds on the possible model uncertainties. Based on such results and the limit cycle analysis, a robust feedback linearization controller is synthesized to eliminate the limit cycle instability.