Diagnostic imaging in a patient with shoulder pain should be used only after a comprehensive clinical evaluation of the shoulder. X-ray and ultrasonography are the basic diagnostic tools; computed tomography and magnetic resonance imaging (MRI) should be used only with certain indications. Ultrasonography and MRI have comparable accuracy for identifying and measuring full-thickness rotator cuff tears, but the accuracy for identifying partial-thickness still needs to be improved. MR arthrography has significantly improved sensitivity and specificity for partial-thickness tears of the cuff. Only ultrasound provides a real-time examination tool during shoulder movements. Moreover, dynamic ultrasonography can assess the contraction patterns of the supraspinatus and infraspinatus muscles, which may improve decision making in the treatment of shoulder diseases. In depicting fatty atrophy of the supraspinatus and infraspinatus muscles, MRI remains the reference standard. MRI should not be used as a diagnostic screening tool in patients with chronic shoulder pain because it does not appear to significantly affect treatment or outcome.