Background: The measurement of the strength of the shoulder muscles is an important element of the overall assessment of patients with various shoulder disorders. However, the clinical utilization of this measurement is dependent on its reproducibility. Objects: To explore the reproducibility of the measurements derived from testing of the isokinetic strength of shoulder muscles in patients with tendinitis of the rotator cuff. Methods: A total of 20 patients with tendinitis of the rotator cuff participated in this study and were assessed twice in 1 week. Isokinetic testing was performed concentrically for shoulder flexors, abductors, and external rotators and eccentrically for the shoulder extensors, adductors, and internal rotators. The relative and absolute reproducibility of the peak torque (PT) and ratios were assessed using intra-class correlation coefficients (ICC), standard error of measurement (SEM), and minimal clinically important difference (MCID), respectively. Results: Overall, high to excellent ICC, clinically acceptable SEM and MCID values were obtained for the PT (ICC: .83-.95, SEM: 1.2%-9%, MCID: 3.4%-25%) and ratios (ICC: .85-.93, SEM: 5.1%-10%, MCID: 14.2%-27.6%). Conclusion: These findings suggest that isokinetic tests may be effectively utilized for the determination of shoulder strength profiles and appropriate position are recommended to perform test without pain in patients with tendinitis of the rotator cuff.