OBJECTIVES: The infant mortality rate is considered an important general indicator both of health conditions as the living conditions and development of a population. By presenting certain sensitivity to various external factors, this rate is directly related to economic and social characteristics. It’s understood that high infant mortality rates are associated with the needs of socioeconomic conditions. Thus, the factors influencing poverty levels (one-dimensional and multidimensional) on Infant Mortality Rate in Brazil are investigated in this paper.  METHODS: So it was considered, in addition to analyzing the impact of one-dimensional poverty, which measures poverty only by per capita income, the impact of multidimensional poverty, whose definition incorporates other dimensions, such as sanitation, housing, information, among others. In order to delineate the models, variables such as per capita income, the Gini index and education level were employed for the Brazilian states from 2001 to 2011. Four models were developed to analyze both individually and together impacts of the explanatory variables on the infant mortality rate. For this purpose, panel data was used on fixed and random effects.  RESULTS: The results showed that decrease the concentration of income and increased educational level, in that order, were the variables which most contributed to the reduction of infant mortality over the period considered. In all, the one-dimensional and multidimensional poverty rates contributed significantly to the reduction of infant mortality rate. However, the ratio of multidimensional poor impact showed a three times higher than the proportion of one-dimensional poor. CONCLUSIONS: The findings indicate that it is essential to the formulation of public policies to reduce income concentration allied to improvements in educational conditions, and reduction of multidimensional poverty rates in order to tackle the problems of poor health, and thus reduce the rate of infant mortality.