Unsafe residential neighborhoods are thought to be a risk factor for childhood obesity by discouraging physical activity while encouraging sedentary behaviors, but findings from cross-sectional studies are susceptible to selection bias (e.g., children who are obese disproportionately reside in unsafe neighborhoods). A literature search was conducted in the Cochrane Library, PubMed, Web of Science, and Google Scholar for articles published until January 2017 that used baseline neighborhood safety levels to predict future weight-related behaviors and body weight status among children aged 17 years and younger. Twenty-two prospective cohort studies conducted in seven countries were identified. The median sample size was 1,104, and the median follow-up was 3.5 years. Sixteen studies used parent- and/or child-reported neighborhood safety measures, and six adopted some objective measures (e.g., county crime rate, interviewers’ block observations). The meta-analysis found that living in unsafe neighborhoods was associated with a reduction in children’s physical activity by 0.13 hours/week. Living in unsafe neighborhoods predicted a trivial (but statistically significant) gain in BMI but no change in childhood overweight/obesity risk. Current research indicates a limited influence of neighborhood safety on childhood obesity; this finding could be partially due to measurement problems. Future longitudinal studies should adopt validated neighborhood safety measures.