Abstract Objectives Preterm delivery (<37 weeks gestation) is the largest cause of child mortality worldwide. Marriage and pregnancy during adolescence have been associated with an increased risk of preterm delivery. We investigate independent associations of age at marriage and age at first pregnancy with preterm delivery in a cohort of women from rural lowland Nepal. Methods We analyzed data from 17 974 women in the Low Birth Weight South Asia Trial. Logistic regression models tested associations of age at marriage and age at first pregnancy with preterm delivery, for primigravida (n = 6 243) and multigravida (n = 11 731) women. Models were adjusted for maternal education, maternal caste, and household asset score. Results Ninety percent of participants had married at <18 years and 58% had their first pregnancy at <18 years. 20% of participants delivered preterm. Primigravida participants married at ≤14 years had higher odds of preterm delivery than those married ≥18 years, when adjusting for study design (adjusted odds ratio (aOR) 1.45, 95% CI: 1.15–1.83), confounders (aOR 1.28: 1.01–1.62) and confounders + age at pregnancy (aOR 1.29: 1.00–1.68). Associations were insignificant for multigravida women. No significant associations were observed between age at first pregnancy and preterm delivery. Discussion In this population, early marriage, rather than pregnancy, is a risk factor for preterm delivery. We hypothesize that psychological stress, a driver of preterm delivery which is increased among those marrying young, rather than physiological immaturity, drives this association. Further research into the psychological consequences of child marriage in Nepal is needed.