International guidance on HIV and infant feeding has evolved over the last decade. In response to these changes, we designed, implemented, and evaluated an education and counseling program for new mothers in Harare, Zimbabwe. The program was implemented within the ZVITAMBO trial, in which 14,110 mother-baby pairs were enrolled within 96 h of delivery and were followed at 6 wk, 3 mo, and 3-mo intervals. Mothers were tested for HIV at delivery but were not required to learn their test results. Infant feeding patterns were determined using data provided up to 3 mo. Formative research was undertaken to guide the design of the program that included group education, individual counseling, videos, and brochures. The program was introduced over a 2-mo period: 11,362, 1311, and 1437 women were enrolled into the trial before, during, and after this period. Exclusive breast-feeding was recommended for mothers of unknown or negative HIV status, and for HIV-positive mothers who chose to breast-feed. A questionnaire assessing HIV knowledge and exposure to the program was administered to 1996 mothers enrolling after the program was initiated. HIV knowledge improved with increasing exposure to the program. Mothers who enrolled when the program was being fully implemented were 70% more likely to learn their HIV status early (<3 mo) and 8.4 times more likely to exclusively breast-feed than mothers who enrolled before the program began. Formative research aided in the design of a culturally sensitive intervention. The intervention increased relevant knowledge and improved feeding practices among women who primarily did not know their HIV status.