Abstarct: Today, robots are successfully applied for high repetitive simple manufacturing tasks that require little interaction of the robot with the environment. It is desirable to extend the capabilities of robots, and apply them in more complex cases, which generally require faster and more accurate motions and more interaction with the environment. This in turn requires the robot to deal with more complex dynamics, which has to be addressed in the controller design. Basically, until recently the main policies of robotic design was baxsed on their high rigidity. However, experience shown that all is not rigidity which obtains by heavy cost of construction. Ignoring flexibility in the controller design, causes instability or at least nominal performance degradations in the under control system. In this thesis, we first review the issue of flexibility in the joints, their problems and some solutions for them. Since, the previous works are baxsed on torque control strategy in the joint space, thus, the first novelty sets has been devoted to extension of control strategies from joint space to task space. For this purpose, composite, nonlinear Proportional-Integral-Derivative, robust backstepping and regressor free adaptive control algorithms are presented. Because of the existence of the actuator electrical dynamics, assuming the availability of joint torques and forces acting on the input can be ruled out. Thus, a second set of initiatives is using the voltage control strategy to controller design in the task-space. At this stage, the simplicity, easy implementation and good performance of the proposed controller is stressed. Also, considering to the new concepts of fractional-order control in robotic systems with flexible joints, the last section of this paper is assigned to design of fractional-order control using voltage control strategy.