The study seeks to estimate the effect of duration, timing and exclusivity of breastfeeding on infant mortality in Kenya. Infant Mortality Rate (IMR) in Kenya remains significantly high at 39 death per 1000 live births (KNBS and ICF Macro, 2014). Optimal breastfeeding is one of the low cost high impact intervention recommended to effect further declines in infant mortality. The purpose of the study was to establish the effect of duration, timing and exclusivity of breastfeeding on infant mortality taking into account other socio-economic and demographic factors known to affect infant mortality. The study utilized secondary data from the Kenya Demographic Health Survey (KDHS 2008/09) with a study population of 4937 births to women of reproductive age (15-49) occurring within three years before the survey. Multivariate logistic regression was the main method of analysis. The dependent variable was occurrence of death during infancy while the independent variables considered for analysis were classified into social-economic and demographic proximate factors. Social-economic factors included wealth index and maternal work status, preceding birth interval, perceived size of infant at birth and sex were included as demographic factors in the multivariate analysis. The life table analysis results shows that infants who were breastfed for more 6 months had a lower mortality risk than those who breastfed for less than 6 months. Logistic regression results established that duration of breastfeeding have strong influence on infant mortality. Infants who were breastfed for more 6 months were less likely to die than those breastfed for 0-6months (OR: 0.020, P<.001). Timing and exclusive breastfeeding were not significantly associated with infant mortality. The study found longer duration of breastfeeding highly beneficial to infant survival and therefore recommends renewed efforts in the implementation of policies and programmes which promote optimal breastfeeding. Future data collection efforts should focus on the completeness of breastfeeding data to facilitate adequate exploration of the role of time to initiation and exclusivity of breastfeeding on infant mortality.