Shoulder pain is a common complaint in elder population and may cause an important functional disability. The aim of this study was to compare the effects of suprascapular nerve block with those of steroid injection in patients with non-specific shoulder pain. A total of 60 patients with shoulder pain lasting for more than four weeks were included in this study. After 1: 1 randomization, patients were included in the steroid group (n = 30) or the suprascapular nerve block group (n = 30). Single injection was performed in both groups. All patients were evaluated before treatment, within one week after treatment, and one month later, in terms of pain, range of motion (ROM) (flexion, abduction, internal and external rotations, and total constant shoulder ROM score), satisfaction, and disability (Pennsylvania shoulder scale function score). Steroid injection was applied at two points in order of lateral and anterior routes. Suprascapular nerve block with lidocaine was applied at the suprascapular notch. No difference was noted in the indicated parameters before the treatment between the groups (p > 0.05). The difference in all follow-up parameters was statistically significant in the assessment periods in both groups (p < 0.05). No method was found to be superior to each other (p > 0.05). No complications occurred in suprascapular nerve block in contrary to steroid injection. We suggest that suprascapular nerve block can be considered the preferred treatment for non-specific shoulder pain because of being as effective as steroid injection with rare side effects. - shoulder pain; corticosteroid injection; suprascapular nerve block (C) 2005 Tohoku University Medical Press.