In spite of various effective intervention programs, the under-5 child mortality rate (U5MR) is still high in Bangladesh. The present paper focused on the levels, trends and determinants of U5MR in Bangladesh utilizing data from Bangladesh Demographic and Health Survey (BDHS), 2007. Differential pattern in U5MR in Bangladesh was examined using life table technique as a bi-variate analysis and Cox proportional hazard model was used to analyze the determinants of U5MR mortality. Data from BDHS (2007) showed that U5MR during the five years preceding the survey was 65 per 1,000 live births, i.e. one in fifteen children born in Bangladesh died before reaching the fifth birthday. The risk of dying in the first month of life (37 per 1,000) was nearly two and a half times greater than in the subsequent 11 months (15 per 1,000). Deaths in the neonatal period accounted for 57 percent of all under-five deaths. In the Cox proportional hazard model analysis, factors such as the place of residence, parent’s education, father’s working status, sources of drinking water, type of toilet facility, wealth status, watching of television, mother’s age, months of breastfeeding, birth interval had significant influence on infant and child mortality. The most significant predictors of neonatal, post-neonatal, infant and child mortality were residence, parent’s education, type of toilet facility, wealth status, watching TV, months of breastfeeding, and birth interval. Despite the improvement of medical technology, child mortality remained alarmingly high, indicating that demographic, socioeconomic, household and environmental conditions must be improved to substantially reduce child mortality in this population.