Background. The effect of exercise on specific impairments and activity limitations in people with hip osteoarthritis (OA) is limited. Objective. The study objective was to evaluate the long-term effect of exercise therapy and patient education on range of motion (ROM), muscle strength, physical fitness, walking capacity, and pain during walking in people with hip OA. Design. This was a secondary outcome analysis of a randomized clinical trial. Setting. The setting was a university hospital. Participants. One hundred nine people with clinically and radiographically evident hip OA were randomly allocated to receive both exercise therapy and patient education (exercise group) or patient education only (control group). Intervention. All participants attended a patient education program consisting of 3 group meetings led by 2 physical therapists. Two other physical therapists were responsible for providing the exercise therapy program, consisting of 2 or 3 weekly sessions of strengthening, functional, and stretching exercises over 12 weeks. Both interventions were conducted at a sports medicine clinic. Measurements. Outcome measures included ROM, isokinetic muscle strength, predicted maximal oxygen consumption determined with the Astrand bicycle ergometer test, and distance and pain during the Six-Minute Walk Test (6MWT). Follow-up assessments were conducted 4, 10, and 29 months after enrollment by 5 physical therapists who were unaware of group allocations. Results. No significant group differences were found for ROM, muscle strength, predicted maximal oxygen consumption, or distance during the 6MWT over the follow-up period, but the exercise group had less pain during the 6MWT than the control group at 10 months (mean difference= –8.5 mm; 95% confidence interval= –16.1, –0.9) and 29 months (mean difference= –9.3 mm; 95% confidence interval= –18.1, –0.6). Limitations. Limitations of the study were reduced statistical power and 53% rate of adherence to the exercise therapy program. Conclusions. The previously described effect of exercise on self-reported function was not reflected by beneficial results for ROM, muscle strength, physical fitness, and walking capacity, but exercise in addition to patient education resulted in less pain during walking in the long term.