The PUMA 560 Industrial Manipulator is presently controlled using a gain-scheduled PID control strategy. The implementation of adaptive control techniques can improve manipulator performance, [20]. One of the principal drawbacks with adaptive control is the difficulty in guaranteeing robustness over the complete operating range. This is particularly important when the plant is expensive and potentially self-damaging. An accurate, tuned, dynamic model, which includes the system’s main dynamics is a great asset to the control engineer. A complete model for the Puma 560 has not previously been developed. The research presented in this thesis derives a tuned and fully tested third order dynamic model for the first three links of the PUMA 560. A mathematical method of specifying the manipulator link structures and inter-relationships is discussed. By treating the manipulator as a connected chain of rigid bodies the Euler-Lagrange method is used to formulate second order dynamic equations and the servo-motor actuators and friction dynamics are incorporated into the model. The resulting third order model is simulated and tested in open and closed loop conditions. The model is then tuned, to provide a better representation of the actual manipulator dynamics, by comparing actual and simulated joint movements for the same actuator voltage input, PID control was implemented in software on each of the model joints independently. Simulated positional joint control resulted in sustained oscillations in joint positions indicating that simple constant gain PID controllers, without dynamic decoupling, were unable to produce satisfactory performance. The PUMA 560 controller board PID parameters were identified for two operating regions of joint 1. These values produced better dynamic performance when implemented in the simulated controller.