Objective: To investigate the diagnostic value of magnetic resonance imaging (MRI) in supraspinatus tendon tears. Methods: 105 patients with shoulder joint lesions were selected as the research subjects and divided into a control group (n=23) and a tear group (n=82) based on the presence of supraspinatus tendon injury. All patients underwent shoulder MRI examination. Using arthroscopy as the “gold standard”, the diagnostic efficacy of MRI was calculated. Patients in the tear group were divided into degeneration group (n=22), partial tear group (n=30), and complete tear group (n=30) according to the degree of injury. The quantitative MRI parameters [diffusion-weighted imaging (DWI) signal intensity, apparent diffusion coefficient (ADC) value, and tendon thickness] between the control group and patients with different degrees of injury were compared, and the correlation with the degree of supraspinatus tendon tear injury, and their diagnostic value for distinguishing tear from simple degeneration were analyzed. Results: The sensitivity, specificity, and accuracy of MRI in diagnosing supraspinatus tendon tears were 95.00%, 95.56%, and 95.24%, respectively. DWI signal intensity and ADC values increased with the aggravation of injury, while tendon thickness peaked in the partial tear group and then decreased (P<0.05). Both DWI signal intensity and ADC values showed a positive correlation with the degree of injury (r=0.891, 0.832, P<0.05). ROC curve analysis showed that the area under the curve (AUC) of ADC values and DWI signal intensity for distinguishing tears from simple degeneration was 0.812 and 0.899, respectively, and the AUC increased to 0.942 when combined. The AUC of tendon thickness was 0.630. Conclusion: MRI has a definite diagnostic effect on supraspinatus tendon tears. DWI signal intensity and ADC values can reflect the severity of injury, and their combination improves the differential diagnostic efficiency. As a morphological indicator, tendon thickness provides supplementary information for injury staging.