Background: This study evaluated the clinical outcome of the arthroscopic capsular release for refractory idiopathic frozen shoulder between diabetic and non-diabetic patients. Patients and Methods: Between August 2020 and April 2022 a prospective study investigated thirty shoulders in 30 patients who had refractory primary frozen and undergone arthroscopic arthrolysis. Assessment was fulfilled using the Constant's shoulder score, Oxford shoulder score, visual analogue scale (VAS) for pain and measuring ranges of motion (ROM) at preoperative, six months postoperatively and at the final follow-up. Results: There were 17 females and 13 males with a mean age of 47.8 (SD 7.5) years. The right shoulder was affected in ten patients and the other 20 had left frozen shoulder. Before surgery, the average time of frozen shoulder was 12.53 (SD 4.2) months. Diabetic patients constituted 56.6%, and mean follow-up was 12.33 (SD 2.9) months. Arthroscopy revealed subacromial adhesions in 73.3% and partial rotator cuff tears in 23%. The Constant's score, Oxford shoulder score, VAS for pain and ROM in all directions significantly increased at the final follow-up, compared to preoperative levels. There were statistically significant differences between diabetic and non-diabetic patients regarding postoperative flexion ROM, postoperative external rotation at 0° abduction, postoperative external rotation with abduction and postoperative internal rotation where non-diabetics showed better ROM and significant improvement in postoperative Constant shoulder score. The mean course of disease after surgery was 3.1 (SD 1.2) and complications represented 6.6%. Conclusions: Arthroscopic release for refractory idiopathic frozen shoulder is an effective procedure. Better results are encountered in non-diabetic patients than diabetic counterparts.