This study examined the moderating effect of low birth weight on the effectiveness of an early intervention program to improve cognitive, language and behavioral outcomes for children prenatally exposed to cocaine. Participants included 293 primarily minority, low SES children who were enrolled in the intervention during their first year and assessed at 36 months. Main effects for intervention and birth weight in the hypothesized direction were found for cognitive, receptive and expressive language ability, and for problem behaviors. There were also significant interaction effects between intervention and birth weight on problem behaviors and prosocial behavior, such that there was an intervention effect on problem behaviors and prosocial behavior for children with low birth weight, but not for children with normal birth weight. These results suggest that, among prenatally cocaine-exposed children, those who have low birth weight may benefit more from intervention than those with normal birth weight. (Contains 2 figures and 2 tables.)