Exercise therapy would appear to be effective at increasing aerobic capacity and muscle strength in patients with rheumatoid arthritis (RA), and no detrimental effects on disease activity or pain compared with controls has been observed. Exercise therapy--at least in the short-term, improves pain, muscular strength and function in elderly people with mild osteoarthritis (OA) of the hip or knee. For the treatment of both OA and RA the knowledge of the optimal type, frequency, duration and intensity of exercise is still limited, but the exercise should not include high-impact loads or high injury risk. Long-term compliance is important in achieving long-term benefits. Supervised classes appears to be as effective as treatments provided on a one-to-one basis, group-based exercise programme thus providing a cost-effective alternative. Adherence to home programmes seems to be lower. Future research should focus on finding optimal type and dose of exercise, ways of optimally maintaining the beneficial effects of exercise therapy over time as well as on the effects of exercise on the long-term progression of the disease and cost-effectiveness of the therapy.