Objective : To investigate whether the frequency of passive motion exercises during the early period after arthroscopic rotator cuff repair affects the range of motion at three months postoperatively. Methods : We retrospectively evaluated 64 patients with rotator cuff tears who underwent arthroscopic repair. Range of motion (ROM) in forward flexion, abduction, external rotation, and internal rotation were assessed preoperatively and at three months postoperatively. Patients were supervised to wear an axillar pillow for six weeks and to commence passive ROM exercise at seven days postoperatively. Passive ROM exercises for 9 hours per week or more were performed in 41 patients (group A), and 3 hours or less of exercises were performed in 23 patients (group B). Results : There were no statistical differences between the two groups in age or gender proportion. The mean differences from preoperative ROM to postoperative ROM were 4° and 17° (group A and B, respectively) in flexion, 5° and 18° in abduction, -14° and -12° in external rotation, and -3 and -2 spinous processes in internal rotation, and there were no differences in all directions between the two groups. Even when subjects were limited to the patients without shoulder contractures preoperatively, there was no difference between the two groups. Conclusion : Early passive motion exercise for long hours after arthroscopic rotator cuff repair did not give a positive effect on early improvement in ROM. Scar formation and adhesion at the subacromial space may influence this outcome.