The majority of manual wheelchair users (MWUs) will inevitably develop some degree of shoulder pain over time. Previous research has suggested a link between the shoulder joint forces associated with the repetition of wheelchair (WC) propulsion and pain. The objective of this work is to present and validate a rigid-body musculoskeletal model of the upper limb for calculation of shoulder joint forces throughout WC propulsion. It is anticipated that when prescribing a WC, the use of a patient-specific computational model will aide in determining an axle placement in which shoulder joint forces are at a minimum, thus potentially delaying or reducing the shoulder pain that so many MWUs experience. During the validation experiment, 3 subjects (2 individuals with paraplegia and one able-bodied individual) propelled a WC at a self-selected speed, during which, kinematics, kinetics, and electromyography (EMG) activity were measured for the contact phase of 10 consecutive push strokes. The measured forces at the push rim and the 3-D propulsion kinematics drove the model, and the computationally calculated muscle activities were compared with the experimental muscle activities, resulting in an average mean absolute error (MAE) of 0.165. Further investigation of the shoulder joint forces throughout propulsion demonstrate the effect of axle placement on the magnitude of these forces. The present work serves to validate the patient-specific upper limb model for use as a prescriptive tool for fitting a subject to their WC. Minimizing joint forces from injury onset may prolong a MWU's pain-free way of life.