This paper presents a comparative study between two control approaches; an Exact FeedForward Linearization controller developed by fuzzy-gain scheduling and sliding mode controller based on Ackermann and Utkin method. For the later one the sliding surface dynamics are determined explicitly without transforming to the sliding mode canonical form. The benefits of this study are demonstrated practically on a well known benchmark control problem, Magnetic ball levitation system (Maglev), and the performances of both controllers are compared. Important control issues such as tracking ability, control effort, steady state error and noise rejection are experimentally investigated.