Using a longitudinal design, the current study explored intimate partner violence perpetration among 1,300 college women within the context of one's history of physical and sexual victimization across 4 years of college. Structural equation modeling indicated that sexual victimization does not predict concurrent use of women's intimate partner violence but does predict subsequent use of women's intimate partner violence during the later years of college. In contrast, physical victimization is associated positively with concurrent use of women's intimate partner violence but is negatively associated with subsequent use of women's intimate partner violence for women. Furthermore, the negative relationship of victimization to subsequent perpetration primarily is due to those with high levels of victimization histories. The present study provides the first model of intimate partner violence within the context of victimization history using longitudinal data. The findings indicate that women's intimate partner violence perpetration is not context-free, but rather is influenced by their own physical and sexual victimization histories.