This study analyzed the effects of physical training programs on heart rate variability, as a measure of sympathovagal balance, in children and adolescents with chronic diseases. Relevant articles were systematically searched in Pubmed, Science Direct, Web of Science, Scopus, Google Scholar and Embase scientific databases. We performed a meta-analysis using an inverse variance heterogeneity model. Effect size calculation was based on the standardized mean differences between pre- and post-intervention assessments, assuring at least a single-group repeated-measures model for each extracted group. Ten studies (252 participants) were included, seven in obese subjects, two in type-1 diabetes, and one in cerebral palsy. When time-domain variables were analyzed, exercise was found to moderately increase RMSSD (SMD=0.478; 95%CI: 0.227 to 0.729; p <0.001), SDNN (SMD=0.367; 95%CI: 0.139 to 0.595; p =0.002) and pNN50 (SMD=0.817; 95%CI: 0.139 to 0.595; p =0.002). As for frequency-domain variables, exercise presented a moderate increasing effect on HF (SMD=0.512; 95%CI: 0.240 to 0.783; p <0.001), a negligible effect for LF (SMD=0.077; 95%CI: -0.259 to 0.412; p <0.001) and a non-significant reduction for LF/HF (SMD=-0.519; 95%CI: -1.162 to 0.124; p =0.114). In conclusion, physical training programs are able to modulate heart rate variability in children and adolescents with chronic diseases, affecting mainly the time-domain variables.