It is important to assess the atrophy of the rotator cuff to better understand shoulder function and pain. Previously, magnetic resonance imaging has been used for the evaluation of atrophy of rotator cuff muscles, which is time consuming. Therefore, a measurement tool requiring little time and easy accessibility is clinically desirable to be used frequently in rehabilitation. Recently, rotator cuff muscles have been evaluated using ultrasonography. However, little is known about the agreement of evaluation in rotator cuff muscles between magnetic resonance imaging and ultrasonography. The purpose of this study was to demonstrate the agreement between the muscle thickness measurements of supraspinatus, infraspinatus, and teres minor muscles by ultrasonography and the cross-sectional area measured by magnetic resonance imaging in the patient with rotator cuff tears. A total of 47 patients with rotator cuff tears were enrolled. There were the 37 small tears, four medium tears, and six large tears, and the involved rotator cuff muscles were the supraspinatus in 37 shoulders, and the supraspinatus and infraspinatus in 10 shoulders. The measuring variables were muscle thickness and cross-sectional area of supraspinatus, infraspinatus, and teres minor muscles by using magnetic resonance imaging. Further, the muscle thickness of the rotator cuff were assessed using ultrasonography. A single regression model was used for demonstrating the agreement between the cross-sectional area measurement by magnetic resonance imaging and the muscle thickness measured using ultrasonography and magnetic resonance imaging of rotator cuff muscles. Additionally, the Bland-Altman plots between magnetic resonance imaging and ultrasonography was analyzed. The cross-sectional area were correlated with the muscle thickness measurement of rotator cuff muscles by magnetic resonance imaging, significantly (supraspinatus: r = 0.84, infraspinatus: ρ = 0.63, teres minor: ρ = 0.61, all p < 0.001). There were significant agreements between the cross-sectional area measured by magnetic resonance imaging and muscle thickness measured by ultrasonography (supraspinatus: r = 0.80, infraspinatus: ρ = 0.78, teres minor: ρ = 0.74, all p < 0.001). Bland-Altman plots revealed significant correlations between the average and the difference of the two measurements in supraspinatus (r = 0.36, p = 0.012), infraspinatus (r = 0.38, p < 0.001), and teres minor (r = 0.42, p < 0.001). These results clarified the proportional bias between MRI and US. This study showed that, similar to magnetic resonance imaging, ultrasonography is a useful tool for assessing muscle atrophy of supraspinatus, infraspinatus, and teres minor muscles.