Purpose: This study was performed to assess the usefulness of non-anatomical repair for irreparable large and massive rotator cuff tears by the arthroscopic margin convergence technique. Materials and Methods: Twenty-two patients were followed up more than 1 year after non-anatomical repair for irreparable large and massive rotator cuff tears using the arthroscopic margin convergence technique. The clinical evaluation was performed according to the KSS score, the UCLA score and the Visual analogue scale (VAS). The measurement of the acromio-humeral distance was performed using the shoulder anterior-posterior radiographs. The measurement of fatty degeneration and the healing status was performed using the shoulder MRI after 6 months. Results: Among twenty-two patients, follow up MRI was performed in eleven cases. Three cases were well healed, four cases were partial healed and another four cases were re-torn. The KSS and UCLA scores had significantly improved from a preoperative average of 45.0${pm}$8.014 and 10.8${pm}$2.302 points to 77.1${pm}$10.151 and 30.0${pm}$1.521 points, respectively, and the pain VAS had decreased from a preoperative average of 7.7${pm}$0.616 points to 3.0${pm}$1.021 points at the last follow up. Less favorable results were obtained when the patient had a grade of fatty degeneration higher than grade 3 on the preoperative MRI. Conclusion: Non-anatomical repair for irreparable large and massive rotator cuff tears by the arthroscopic margin convergence technique showed good functional results. It seems to be one of the effective treatment methods for irreparable large and massive rotator cuff tears.