Aim: To investigate whether a supervised exercise program can reduce or prevent musculoskeletal pain in women with breast cancer undergoing AI therapy. Methods: Twenty participants with breast cancer have thus far been randomised to, and completed either: (a) usual care with advice regarding benefits of regular exercise; or (b) usual care + 12 week supervised gym-based and home exercise program consisting of upper and lower body resistance exercises with self-selected aerobic exercise. Participants accrue 150 mins / week of moderate intensity aerobic exercise at 60-70% HRmax on 5 or more days of each week including 2 supervised sessions / week. Initial exercise intensity is individualized and generally begins at 55% to 60% of HRmax (15 to 30 minutes per session) and progresses to 60% - 70% of HRmax by week 6. Strength training consists of 2 supervised sessions per week. On at least 3 other days, participants perform resistance band exercises at home. The primary outcome measures are pain scores (brief pain inventory; BPI) and grip strength (JAMAR dynamometer (kg)) measured at baseline, 3, 6 and 12 months. Two-way repeated ANOVAs evaluate differences between groups and factors. Results: Mean pain scores increased by one BPI unit between baseline and 12 month follow-up (p>0.05) for participants undertaking the exercise program, compared to control participants whose mean BPI scores increased by five BPI units over the 12 month period. Grip strength measures were significantly different (P<0.001) between groups at each time point with a trend towards improved grip strength between baseline and 6 months in the exercise group. Grip strength decreased in both groups between 6 and 12 months. One patient in each group discontinued AI therapy. Conclusion: A 12 week supervised exercise program combined with home-based exercises may control AI-induced musculoskeletal pain.