This six-month follow-up study aimed to assessing the effect of length of both exclusive breastfeeding and mixed feeding on hemoglobin levels in 150 infants born at the Mutuípe maternity facility in Bahia State, Brazil. Blood tests were performed monthly to assess the hemoglobin levels, at which time data were gathered on food intake. Obstetric and socioeconomic data were also collected. The linear mixed effects model was used to investigate the target associations. Loss to follow-up was 15.3%, and anemia prevalence was 71.7%. An increase of 0.091 g/dL (p = 0.031) in hemoglobin levels was found for each month of exclusive breastfeeding, while a decrease of 0.097 g/dL (p = 0.017) was shown for each month of mixed feeding. In conclusion, inadequate food practices appear to be directly involved in the etiology of iron deficiency anemia during infancy.