This article describes a 20-year program of research on the Nurse Home Visitation Program, a model in which nurses visit mothers beginning during pregnancy and continuing through their children's second birthdays to improve pregnancy outcomes, to promote children's health and development, and to strengthen families' economic self-sufficiency. The results of two randomized trials (one in Elmira, New York, and the second in Memphis, Tennessee) are summarized, and an ongoing trial in Denver, Colorado, is briefly described. Results of the Elmira and Memphis trials suggest the following: The program benefits the neediest families (low-income unmarried women) but provides little benefit for the broader population. Among low-income unmarried women, the program helps reduce rates of childhood injuries and ingestions that may be associated with child abuse and neglect, and helps mothers defer subsequent pregnancies and move into the workforce. Long-term follow-up of families in Elmira indicates that nurse-visited mothers were less likely to abuse or neglect their children or to have rapid successive pregnancies. Having fewer children enabled women to find work, become economically self-sufficient, and eventually avoid substance abuse and criminal behavior. Their children benefitted too. By the time the children were 15 years of age, they had had fewer arrests and convictions, smoked and drank less, and had had fewer sexual partners. The program produced few effects on children's development or on birth outcomes, except for children born to women who smoked cigarettes when they registered during pregnancy. The positive effects of the program on child abuse and injuries to children were most pronounced among mothers who, at registration, had the lowest psychological resources (defined as high levels of mental health symptoms, limited intellectual functioning, and little belief in their control of their lives). Generally, effects in Elmira were of greater magnitude and covered a broader range of outcomes than in Memphis, perhaps because of differences between the populations studied, community contexts, or a higher rate of turnover among home visitors in Memphis than in Elmira. The article concludes that the use of nurses as home visitors is key; that services should be targeted to the neediest populations, rather than being offered on a universal basis; that clinically tested methods of changing health and behavioral risks should be incorporated into program protocols; and that services must be implemented with fidelity to the model tested if program benefits found in scientifically controlled studies are to be reproduced as the program is replicated in new communities.