Objectives: Our objectives were to characterize the stability of pregnancy intention and to examine whether stability is associated with the timing of prenatal care initiation, smoking during pregnancy, and breastfeeding. Methods: We use a sample of women from the National Longitudinal Survey of Youth (NLSY) for whom information on pregnancy intention was collected both during pregnancy and after delivery. In bivariate analyses we compare outcomes and characteristics of women whose pregnancy intention changed between the prenatal and postpartum periods. With multivariate methods, we analyze the correlates of switching pregnancy intention as well as the association between switching and maternal behaviors. Results: Women whose pregnancy intention changes between the two assessments are similar in marital status and socioeconomic background to those who report both during pregnancy and after delivery that the pregnancy is unintended. Disagreement during pregnancy between the parents' pregnancy intentions is the most important predictor of instability in the mother's pregnancy intention. Effects of unintended pregnancy on the timing of initiation of prenatal care, smoking during pregnancy, and breastfeeding based on reports after delivery are smaller than those based on reports during pregnancy, although differences are not statistically significant. Adverse effects of unintended pregnancy are greater when pregnancies reported by the mother to be unintended at either assessment are combined into a single category for unintended pregnancy. Conclusion: Unstable pregnancy intention may be a marker for adverse maternal behaviors related to infant health.