People with osteoarthritis (OA) suffer from joint degeneration and pain as well as difficulty performing daily activities. Joint contact forces (JCF) are important for understanding individual joint loading, however, these contact force cannot be directly measured without instrumented implants. Musculoskeletal modeling is a tool for estimating JCF without the need for surgery. The results from these models can be very different due to different approaches used in the development of a model that was used for simulation. Therefore, the first purpose of this study was to develop and validate a musculoskeletal model in which lower extremity JCF were calculated at the hip, knee, and ankle in 10 participants with hip OA (H-OA) and 10 healthy control participants using OpenSim 4.0 [simtk.org, 23]. The generic gait2392 model was scaled to participant demographics, then the inverse kinematics (IK) solution and kinetic data were input into the Residual Reduction Algorithm (RRA) to reduce modeling errors. Kinematic solutions from RRA were used in the Computed Muscle Control (CMC) tool to compute muscle forces, then JCF were estimated using the Joint Reaction Analysis tool. Validation included JCF comparisons to published data of similar participant samples during level walking, and movement simulation quality was assessed with residual forces and moments applied at the pelvis, joint reserve actuators, and kinematic tracking errors. The computed JCFs were similar to the overall trends of published JCF results from similar participant samples, however the values of the computed JCFs were anywhere from 0.5 times body weight (BW) to 3BW larger than those in published studies. Simulation quality assessment resulted in low residual forces and moments, and low tracking errors. Most of the reserve actuators were small as well, besides pelvis rotation and hip rotation. The computed JCF were then used in the second portion of this study to determine the effect of group and side on JCF during both the weight acceptance and push-off phases of level walking. It was determined that there was a significant difference in the knee and ankle JCF during the weight acceptance portion of stance phase and at all joints during the push-off phase when comparing the H-OA and control groups on the affected limb. A significant interaction between group and limb was found for the peak hip JCF timing (% stance) during the push-off portion of the stance phase (p=0.009). These results demonstrate that H-OA participants experience an earlier peak hip JCF during propulsion on their affected limb. Based on previous research in OA that has examined spatiotemporal measures, this finding suggests that H-OA participants may use step or stride length changes as a strategy to decrease or limit pain and loading on the affected limb. Knowledge of potential JCF differences in H-OA participants, such as timing of the peaks in either portion of the stance phase, could provide useful insight to clinicians and therapists to make decisions on how to proceed with treatment or rehabilitation programs.