This meta-analysis aimed to evaluate the association between childhood and adolescent obesity and depression. We systematically searched PubMed, PsycInfo, EMBASE and Science Direct for studies that compared prevalence of depression and depressive symptoms in normal weight and obese children and adolescents. Observational studies were included if they reported body mass index and assessed depression by validated instruments or diagnostic interviews. Quality assessment was performed using the Newcastle-Ottawa scale. We used the random-effect model to calculate the pooled odds ratios, standard mean differences (SMDs) and subgroup analysis. Findings for a total of 51,272 participants were pooled across 18 studies and examined. Our analyses demonstrated a positive association between childhood and adolescent obesity and depression (pooled odds ratio = 1.34, 95% confidence interval [CI]: 1.1-1.64, p = 0.005) and more severe depressive symptoms (SMD = 0.23, 95% CI: 0.025-0.44, p = 0.028) in the obese groups. Overweight subjects were not more likely to have either depression (pooled odds ratio = 1.16, 95% CI: 0.93-1.44, p = 0.19) or depressive symptoms (SMD = 0, 95% CI: -0.101 to 0.102, p = 0.997). Non-Western and female obese subjects were significantly more likely to have depression and severe depressive symptoms (p < 0.05). In conclusion, obese children and adolescents are more likely to suffer from depression and depressive symptoms, with women and non-Western people at higher risk.