The process of divorce is usually lengthy and hazardous, and can start quarrels that can lead to the abuse of women and their children. This study examines the effects of divorce on neonatal and postneonatal mortality of babies born before and after divorce in Teknaf, a remote area of Bangladesh. The longitudinal demographic surveillance system (DSS) followed 1762 Muslim marriages in 1982–83 for 5 years to record divorce, deaths of spouse, emigration and births. It recorded 2696 live births during the follow-up period, and their survival status during infancy. Logistic regression models were used to estimate the effect of divorce on neonatal and postneonatal mortality, controlling for maternal age at birth, parity, sex of the child and household economic status. The odds of neonatal and postneonatal deaths among babies born after divorce or less than 12 months before mothers were divorced were more than double the odds of those born to mothers of intact marriages. The odds of postneonatal deaths were two times higher among babies born more than 12 months before divorce happens than their peers. The high mortality of infants born before and after mothers were divorced may reflect how abusive marriage and divorce increase the vulnerability of women and children in rural Bangladesh. Divorce and abuse of women are difficult and intractable social and health problems that must be addressed.