We evaluated the efficacy/effectiveness of community-based interventions to decrease neonatal mortality. A systematic review of randomized controlled trials, cluster randomized trials and cohort studies of interventions on pregnant women, neonates (up to 28 days after birth) or both was made. Thirty four studies were evaluated (n=844,989): 20 in pregnant women (n=406,172), 6 in neonates (n=24,994), and 8 in both (n=413,823). Risk of bias was generally low. There was heterogeneity among interventions. Interventions such as maternal health education and maternal and neonatal home care were associated to a decrease in neonatal mortality in half of the 6 studies of each group. Supplementation with multiple micronutrients, kangaroo mother care, and maternal supplementation with vitamin A did not decrease neonatal mortality. A few heterogeneous community-based interventions demonstrated a decrease in neonatal mortality.