Objective: To investigate the mid-term effectiveness of arthroscopic Bankart repair for recurrent anterior shoulder dislocation. Methods: The clinical data of 107 patients with recurrent anterior shoulder dislocation who met the inclusion criteria between January 2017 and June 2021 was retrospectively analyzed, and all patients underwent arthroscopic Bankart repair. There were 88 males and 19 females. The age of the primary dislocation ranged from 13 to 48 years (mean, 23.3 years). The number of preoperative dislocations was 2-160 times (median, 7 times). The duration of preoperative instability was 0.2-240.0 months (median, 36.0 months). The mean age at operation was 28.2 years (range, 16-61 years). There were 43 cases of left shoulder and 64 cases of right shoulder. The proportion of glenoid defects in 63 patients was 1.7%-16.1% (mean, 8.1%). MRI showed that none of the patients had rotator cuff tears or shoulder stiffness. The CT three-dimensional reconstruction was performed at 1 day after operation to evaluate the distribution of implanted anchors and the occurrence of glenoid split fracture and whether there were nails pullout at the implant site. The postoperative complications were observed, and the pain and function of the shoulder were evaluated by visual analogue scale (VAS) score, Rowe score, Constant-Murley score, and American Shoulder and Elbow Surgeons (ASES) score. The recurrence of instability, the results of apprehension test, the number of patients who returned to preoperative sports level, and the satisfaction rate of patients were recorded. Results: All patients were successfully operated and were followed up 20-73 months (mean, 41.5 months). All incisions healed by first intention. The CT three-dimensional reconstruction at 1 day after operation showed that the anchors were located at the 2 : 00-5 : 30 positions of the glenoid, and there was no glenoid split fracture or nails pullout at the implant site. At last follow-up, VAS score was significantly lower than that before operation, and Rowe score, Constant-Murley score, and ASES score were significantly higher than those before operation ( P<0.05). Seven patients (6.5%) had recurrence of anterior shoulder dislocation at 23-55 months (mean, 39.9 months) after operation, including 6 cases of dislocation and 1 case of subluxation. At last follow-up, 51 patients (47.7%) returned to preoperative sports level, and 11 patients (10.3%) had a positive apprehension test. The patients' satisfaction rate was 90.7% (97/107). Among the 10 patients who were not satisfied with the surgical effectiveness, 7 patients had postoperative recurrence of instability, and 3 patients felt that they did not return to preoperative sports level. Conclusion: Arthroscopic Bankart repair has good mid-term effectiveness in patients with recurrent anterior shoulder dislocations, minimal or no glenohumeral bone defects and low sports need.