Music therapy has been explored as a non-pharmacological intervention for anxiety in critically ill patients, but its effectiveness remains unclear. This meta-analysis synthesized evidence from five randomized controlled trials (RCTs) involving 325 participants. A comprehensive literature search across PubMed, Embase, and Cochrane databases ensured the inclusion of high-quality studies published between 2010 and 2025. Music therapy sessions ranged from 15 to 30 min. Anxiety was assessed using validated measures, and data were analyzed with a random-effects model to calculate the mean difference (MD) in anxiety scores with 95 % confidence intervals (CI). Heterogeneity was evaluated using the I2 statistic, and sensitivity analyses were conducted to address potential bias. The pooled analysis found an MD of -1.11 (95 % CI: -12.97 to 10.75, p = 0.85; I2 = 99 %), showing no significant anxiety reduction. Heterogeneity was substantial. A sensitivity analysis, excluding two influential studies, reduced heterogeneity (I2 = 30 %) but still showed no significant effect (MD = 0.69, 95 % CI: -2.15 to 3.53, p = 0.63). A funnel plot suggested possible publication bias, but the limited number of studies hindered formal assessment. Overall, current evidence does not support music therapy as an effective stand-alone intervention for anxiety reduction in critically ill patients. Further research with standardized protocols and larger sample sizes is necessary to clarify its role in critical care settings.