Many individuals who have suffered from a spinal cord injury require assistance to perform activities of daily living, and this population considers regaining upper limb function as a top priority to restore independence. Robotic exoskeletons and functional electrical stimulation are two technologies that can provide some amount of aid in these cases, but each technology alone has limitations that keeps it from providing meaningful assistance for daily activities. Combining these two technologies could counter these limitations by allowing functional electrical stimulation to provide large amounts of the general power requirements, while an exoskeleton can fine tune movements, allowing for meaningful assistance. However, this combination also raises a new challenge – effectively distributing control effort between the two sources with differing time delays while maintaining high accuracy in coordinated movements. This thesis presents a model predictive control-based hybrid controller which utilizes the cost function to achieve this goal. This hybrid controller is implemented and tested in both single and multi-joint movements in its ability perform trajectory tracking tasks. Findings from studies with healthy participants indicate that the hybrid controller is able to reduce exoskeleton control effort compared to an exoskeleton acting alone yet maintain high accuracy in controller implementations with one and two joints, and simulations in four joint movements show ideal controller behavior, outlining potential capabilities in highly complex movements.