This paper presents a robust control law for a redundant robotic arm that manipulates an unknown load attached to its end effector. As it is considered that the robot manipulator may interact with the environment, the control law is expressed in terms of the operational space variables. An extended dynamic model of the robot, considering the uncertain load and intentional simplifications in the kinematic model, is presented for the control system design. The objective of the work is to develop, based in a sliding control mode, a robust motion control law featuring robustness in face of the above mentioned model uncertainties and simplifications, and to prove, with Lyapunov’s theory, the stability of the closed loop system. The kinematic redundancy of the manipulator is exploited to enhance the performance. It is shown, through simulations done on a 7-DOF KUKA LBR iiwa 14 R820 robot model, that the errors remain bounded when the box is following a trajectory even with different values of loads.