In recent years, increasing interest has been devoted to the physiological basis of self and other-oriented compassion. Heart rate variability (HRV) represents a promising candidate for such a role, given its association with soothing emotions and context appropriate prefrontal inhibitory control over threat-defensive responses. The aim of this study was to meta-analyze available studies on the association between compassion and HRV. Random-effect models were used. The analysis performed on sixteen studies that met inclusion criteria, yielded a significant association with a medium effect size (g = .54 95% CI [.24, .84], p < .0001). Results were not influenced by publication bias. After an extreme outlier’s exclusion, the size of the association was still larger in studies that used time or frequency-domain indices of vagally-mediated HRV compared to those that used peak to trough estimates of respiratory sinus arrhythmia. Results are limited by the small number of studies included in the meta-analysis (n = 16) and are discussed in terms of indications for future research, given that existing data are highly heterogeneous and of poor methodological rigor.