Abstarct: In this thesis, a new mathematical model of robotic arm will be presented. Proposed control laws baxsed on this model is of adaptive fuzzy sliding mode type in task-space which is able to cope with different uncertainties such as parametric uncertainty, robot lixnks friction, external disturbance, kinematics and Jacobian matrix uncertainties and unmodeled dynamic using estimation of uncertainties bound. The control laws are been baxsed on torque control strategy. In this approach, torque can not be applied directly to robot lixnks. In other hand, we are in face to very complex equations and more computation in torque control strategy. Thus, in this research, we use the novel voltage control approach where achieves better and faster performance and less computationl. Also in aforementioned strategy, we do not deal with dynamical model of robot. It is an important benefit in comparision with traditional torque control strategy. In this thesis, two control schemes baxsed on voltage control strategy and a control scheme baxsed on torque control strategy is presented. In proposed schemes, firstly equations related to voltage and torque is turned to known form and then control laws are proposed baxsed on them. Stability proof of presented methods is baxsed on Lyapunov function. In the end, a comparision between adaptive fuzzy sliding mode using torque control strategy and voltage control strategy are performed. According to this comparision, tracking error of end-point of robot in voltage control approach is less than torque control approach. In presented control law baxsed on proposed approach, it is not necessary to measure the velocity. This is while, in proposed control law baxsed on conventional torque control strategy, we need to measure the velocity plus to measure the position. Less computation in voltage control approach is another result which is important in point views of practical applications.