Heart rate variability biofeedback (HRVB) has been used for a number of years to treat depressive symptoms, a common mental health issue, which is often comorbid with other psychopathological and medical conditions. The aim of the present meta-analysis is to test whether and to what extent HRVB is effective in reducing depressive symptoms in adult patients. We conducted a literature search on Pubmed, ProQuest, Ovid PsycInfo, and Embase up to October 2020, and identified 721 studies. Fourteen studies were included in the meta-analysis. Three meta-regressions were also performed to further test whether publication year, the questionnaire used to assess depressive symptoms, or the interval of time between T0 and T1 moderated the effect of HRVB. Overall, we analysed 14 RCTs with a total of 794 participants. The random effect analysis yielded a medium mean effect size g = 0.38 [95% CI = 0.16, 0.60; 95% PI =  - 0.19, 0.96], z = 3.44, p = 0.0006. The total heterogeneity was significant, Q = 23.49, p = 0.03, I = 45%, which suggested a moderate variance among the included studies. The year of publication (χ(1) = 4.08, p = 0.04) and the questionnaire used to assess symptoms (χ(4) = 12.65, p = 0.01) significantly moderated the effect of the interventions and reduced heterogeneity. Overall, results showed that HRVB improves depressive symptoms in several psychophysiological conditions in adult samples and should be considered as a valid technique to increase psychological well-being.