This paper modeled the proximate determinants of infant survival using the National Family Health Survey data on 11,500 women from the most populous Indian state Uttar Pradesh in the period 1982–1992. A methodological framework was developed for analyzing the inter-relationships between high fertility and infant mortality, gender differences in mortality, and for modeling the effects of health care and family planning variables. Probit models were estimated by maximum likelihood taking into account simultaneity of regressors and unobserved household differences. The proximate determinants of infant survival included maternal education and age at first birth, birth interval, the number of children before family planning was first used, maternal tetanus vaccination, and child's vaccinations. Indicator variables for a boy (girl) born at a birth order higher than the “ideal” number showed that unwanted births exacerbated female mortality