We aimed to compare the clinical efficacy of a suprascapular nerve block (SSNB) versus subacromial injection (SA) for outpatient treatment of patients with symptomatic rotator cuff tears in a double-blinded, randomized controlled trial using sealed-envelope randomization. A total of 42 participants with symptomatic partial- and full-thickness rotator cuff tears quantified by ultrasound or magnetic resonance imaging received either an ultrasound-guided SSNB or SA. The primary outcome measure was shoulder function measured by the modified Constant-Murley (CM) score and the secondary outcome was the pain score measured by a visual analog scale at 2, 6, and 12 weeks after injection. We analyzed 43 shoulders (27 in male patients, 62.2%). The mean age was 65.2 years (standard deviation [SD], 11.9 years). Of the shoulders, 22 (51.2%) underwent SAs and 21 (48.8%) underwent SSNBs. Continuous variables were analyzed by an independent t test (2 tailed), and nominal data were analyzed by the Fisher exact test (1 sided). At 6 weeks, the mean change from the baseline CM score was significantly higher in the SSNB group than in the SA group (14.3 [SD, 18.1] vs 3.0 [SD, 12.8]; P = .048). At 12 weeks' follow-up, the SSNB group had a significantly higher CM score than the SA group (57.6 [SD, 10] vs 44.6 [SD, 16]; P = .023) and greater improvement from the baseline CM score (23.4 [SD, 17.5] vs 7.8 [SD, 16.5]; P = .014). At 12 weeks, the visual analog scale score was significantly better in the SSNB group than in the SA cohort (9.9 [SD, 3.3] vs 7.3 [SD, 4.3]; P = .03). This study demonstrates that an SSNB resulted in better pain and functional results than an SA at 6 and 12 weeks for symptomatic rotator cuff tears.