This paper investigated the determinants of unhealthy BMI categories, namely underweight, overweight, and obesity, for urban and rural women of childbearing age in Bangladesh. Multilevel nested logistic models were estimated using data from multiple waves (2004–18) of the Demographic and Health Surveys of Bangladesh. After controlling for community and regional effects, the underweight status showed an inverse relationship with years of education for urban and rural women. Conversely, the fitted model estimates revealed a positive association between overweight risk and education for both urban and rural women populations. However, the association between obesity and education was significant for only urban women. Employment status appeared to be negatively associated with the risks of being overweight and obese in both population groups. Moreover, the women were found less likely to be underweight and more likely to be overweight and obese when their household socio-economic status improved. Study findings have the potential to inform a wide range of interventions, policies and programs that simultaneously address the adverse consequences of underweight, overweight, and/obese population health conditions in rural and urban Bangladesh.