The brain adjusts activities of muscles to control joint positions with a motion control policy. To identify the motion control policy, mechanical action of each muscle, which is determined by spatial relationship between the muscle and relevant bones, is the most important factor. The relationship determines the moment arm of the muscle and thereby provides the basis for allotting a functional role to each muscle in controlling the joint position. Therefore, it is critical to analyze the mechanical property of each muscle on the joint. Based on this reasoning, we focused on the mechanical actions of wrist prime movers on the wrist joint. We supposed this mechanical property of muscles on the wrist provide the motive to analyze the control policy between multiple muscles in controlling the joint. To do this, we first designed a musculoskeletal model of the wrist joint with precise paths of relevant muscles relative to the bones based on anatomical data. To evaluate our model, we measured the pulling direction (PullD) of four wrist prime movers on the wrist joint by electrically stimulating each muscle with a pair of wire electrodes inserted into the muscle. A PullD is measured as direction of the evoked movements of each muscle on the wrist joint. We evaluated the proposed model by comparing the measured PullD and the simulated PullD. Our model can precisely estimate the change of mechanical parameters such as muscle length in controlling the joint, which cannot be measured directly in the experiment. We are going to discuss the relation of these mechanical parameters and the control policy of multiple muscles for control of joint position.