Using guidelines of the Meta-analysis of Observational Studies in Epidemiology Group, we systematically reviewed the literature on neonatal jaundice (unconjugated hyperbilirubinemia) and Autism Spectrum Disorder (ASD) in term and preterm infants. Thirteen studies were included in a meta-analysis. Most used retrospective matched case-control designs. There was significant heterogeneity (Q = 31, p = 0.002) and no evidence of publication bias (p = 0.12). Overall, jaundice, assessed by total serum bilirubin (TSB), was associated with ASD (OR, 1.43, 95% CI 1.22-1.67, random effect model). This association was not found in preterms (OR 0.7, 95% CI 0.38-1.02) but deserves further investigation since other measures of bilirubin such as unbound unconjugated bilirubin may be better predictors of neurotoxicity than TSB in preterms.