This paper proposes the application of fuzzy and neuro-fuzzy techniques to the control of an open loop unstable magnetic suspension system using genetic algorithms (GA's) as a tuning algorithm. A Mamdani type fuzzy PID controller is investigated. The proposed controller employs a twostage control structure, where the fuzzy PI controller reasonably aims to suppressing steady state errors and the fuzzy PD controller provides stabilizing action. The membership functions' parameters for both controllers are determined by the GA's. A Takagi-Sugeno-Kang (TSK) controller using parallel distributed compensation (PDC) approach is presented. The proposed controller, which is nonlinear, is fuzzy blending of individual piecewise linear controllers designed using pole placement technique. A neuro-fuzzy controller is proposed. Reinforcement learning along with GA's was used to train the neuro-fuzzy controller.