Although umbilical cord infection contributes to neonatal mortality and morbidity and risk can be reduced with topical chlorhexidine, behavioral or other factors associated with cord infection in low-resource settings have not been examined. Data on potential risk factors for omphalitis were collected during a community-based, umbilical cord care trial in Nepal during 2002-2005. Newborns were evaluated in the home for signs of umbilical cord infection (pus, redness, and swelling). Omphalitis was defined as either pus discharge with erythema of the abdominal skin or severe redness (>2 cm extension from the cord stump) with or without pus. Multivariable regression modeling was used to examine associations between omphalitis and maternal, newborn, and household variables. Omphalitis was identified in 954 of 17,198 newborns (5.5%). Infection risk was 29% and 62% higher in infants receiving topical cord applications of mustard oil and other potentially unclean substances, respectively. Skin-to-skin contact (relative risk (RR) = 0.64, 95% confidence interval (CI): 0.43, 0.95) and hand washing by birth attendants (RR = 0.73, 95% CI: 0.64, 0.84) and caretakers (RR = 0.76, 95% CI: 0.60, 0.95) were associated with fewer infections. In this community, unhygienic newborn-care practices lead to continued high risk for omphalitis. In addition to topical antiseptics, simple, low-cost interventions such as hand washing, skin-to-skin contact, and avoiding unclean cord applications should be promoted by community-based health workers.