Objective To investigate the early effectiveness of the limited unique coracoid osteotomy suture button fixation Latarjet (LU-tarjet)-congruent-arc (CA) technique (LU-tarjet-CA) in treating recurrent shoulder dislocations with huge glenoid defect. Methods The clinical data of 12 patients with recurrent shoulder dislocation and huge glenoid defect who met the selection criteria and treated with arthroscopic LU-tarjet-CA between January 2021 and December 2023 were retrospectively analyzed. The cohort included 8 males and 4 females, aged 20-40 years with an average age of 30.4 years. The range of glenoid bone loss was 30%-40%, with an average of 35.5%. The time from symptom onset to hospital admission ranged from 1 to 36 months, with an average of 18.5 months. The University of California Los Angeles (UCLA) score, American Association for Shoulder and Elbow Surgery (ASES) score, Walch-Duplay score, and Rowe score were used to evaluate shoulder function preoperatively and at 3, 6, and 12 months postoperatively. CT three-dimensional (3D) reconstruction was used to assess coracoid healing and plasticity at 3, 6, and 12 months postoperatively. Subjective satisfaction of patient was recorded at last follow-up. Results All incisions healed by first intention, with no incision infection or nerve injury. All 12 patients were followed up 12 months. One patient developed Propionibacterium acnes infection within the joint postoperatively and recovered after initial arthroscopic debridement and anti-inflammatory treatment. At 3 months after operation, CT 3D-reconstruction showed 1 case of complete coracoid absorption; neither of these two patients experienced redislocation. The remaining patients exhibited partial coracoid absorption but displayed local reshaping, filling the preoperative defect area, and bony fusion between the coracoid and the glenoid. At last follow-up, 9 patients (75%) were very satisfied with the outcome, and 3 patients (25%) were satisfied; the satisfied patients experienced postoperative shoulder stiffness caused by suboptimal functional exercise but did not have impaired daily life activities. The UCLA score, ASES score, Walch-Duplay score, and Rowe score at 3, 6, and 12 months postoperatively were significantly better than preoperative scores, and each score improved further over time postoperatively, with significant differences between different time points (P<0.05). Conclusion The arthroscopic LU-tarjet-CA technique for treating recurrent shoulder dislocations with huge glenoid defect can achieve the surgical objective of bony blockade and filling bone defects to prevent shoulder dislocation, thereby improving patients’ quality of life and shoulder joint function and stability.