District level trend of infant mortality rate (IMR) per thousand live births in Bangladesh influenced by some assorted form of socio-demographic determinants such as individual, household and community level factors. This paper examines the trend and annual rate of reduction from 1998-2007 time periods and correlates causal factors based on different data from Statistical Yearbook of Bangladesh 2008 and Sample Vital Registration System 2007. Seven explanatory variables are considered and the log-log specified ordinary least square and simultaneous quantile regression models are employed to investigate and compare the stochastic impacts of these predictors on changing infant mortality. Infant immunization is the most effective factor that reduces infant mortality especially at lower quantile districts. Most notably, lower poverty line implies increasing trend with upper quantile, indicates that districts with low infant mortality rate has low effect for any positive rate of change of it. The least square as well as simultaneous quantile regression result disclose that share of population lived in electricity accessed houses, road density, no. of female per family planning personnel has potential and statistically significant impacts on infant mortality rate that is -0.25%, -0.22% and -0.58% respectively. Likewise, infant mortality decreased with the increased percentage of household having television by 0.08%, on average. As infant mortality is an outcome from a variety of socio-economic disparity; reduction strategy should address the degree of severity of the risk factors on infant mortality, prioritizing the most effective reducing factors such as infant immunization and controlled population growth rate as well.