Energy saving, low robot mass to carried mass ratio, more ability to work in various environments, easier delivery of parts and lower production costs in flexible robots make these robots more attractive than rigid robots to many researchers and industries. But due to nonlinearities in flexible robot system and high vibration in operation points and also more sensitivity against external disturbances, control of these robots is more difficult and complex. In this paper a controller for a flexible link manipulator based on fractional calculus is practically implemented. At first the dynamic model of a single flexible-link robot is introduced. Then various controllers such as fuzzy control, PID control, and fractional order PID torque control are practically implemented on a single flexible-link robot made in laboratory, and then the performance of each controllers in decreasing of arm vibration in final desired point and tracking error reduction are investigated. Further, to compare the robustness of the designed controllers, a same constant disturbance is applied to all controllers and their performance are compared. Finally, the simulation results and experimental results show that the fractional order PID torque controller has the best results among the implemented controllers.