The PUMA 560 Industrial Manipulator is presently controlled using a PID control strategy Robot manipulators are highly coupled, nonlinear mechanical systems designed to perform specific tasks. It is the function of any control algorithm to compute the input voltages or torques needed to follow a desired trajectory. The PID controller is detuned, so as to cater for variations in system behaviour. Thus, the performance of such a control algorithm is poor over the entire operating range of the robot and the need for more complex control strategies is clear. The research presented m this thesis derives a third order comprehensive dynamic model for the three primary robot joints, using the Euler-Lagrange formulation for the equations of motion. A simulation package is designed to model this dynamic system. Next, a wide range of different techniques are investigated in a simulation environment, to observe their performance on the computer model. These control algorithms range from Fixed Parameter techniques to Adaptive strategies and Feedforward routines. A set of performance criteria can be used to evaluate these techniques, and the best algonthm from each section is chosen. Using the results of an identification performed on the robot, each of these control methods is applied to the resulting tune varying model. The results here are used to determine the optimal control strategy for manipulator use. Also in this thesis, a new hardware structure is designed and implemented. This structure is capable of implementing complex control routines with adequately low sample periods. The design uses advanced digital signal processors, which can perform arithmetic operations quickly.