This paper introduces a novel wearable full wrist exoskeleton designed for the alleviation of tremor in patients suffering from Parkinson’s Disease and Essential Tremor. The design introduces a structure to provide full observation of wrist kinematics as well as actuation in wrist flexion/extension and radial/ulnar deviation. To examine the feasibility of the design, the coupled dynamics of the device and the forearm is modeled via a general multibody framework. The dynamic analysis considers human motion, wrist stiffness, and tremor dynamics. The analysis of the model reveals that the identification of the wrist kinematics is indispensable for the controller design. Nonlinear regression based on the Levenberg-Marquardt algorithm has been applied to estimate the unknown parameters in a kinematic structural function designed to approximate the wrist kinematics, which leads to the construction of the control system framework. Finally, several simulation cases are demonstrated to conclude the study.