In this study, neuro-fuzzy control of Quanser flexible link system is presented. Since flexible link has an under actuated nature, the PD-like and importance-based neuro-fuzzy (NF) controllers have to control the hub position so that it can attenuate the tip deflections. After simulation, both controllers are implemented on the Quanser flexible link system. As all different mathematical models of the flexible link deal with uncertainties, a comparison is made on the robustness of these controllers in addition to their performance.