Human life history theory describes how resources are allocated among conflicting life tasks, including trade-offs concerning reproduction. The current research investigates the unique importance of environmental unpredictability in childhood in association with romantic attachment, and explores whether objective or subjective measures of environmental risk are more informative for testing life history hypotheses. We hypothesize that (1) unpredictability in childhood will be associated with greater anxious attachment, (2) anxious attachment will be associated with intimate partner violence (IPV) perpetration, and (3) anxious attachment will mediate the relationship between unpredictability in childhood and IPV perpetration. In two studies (total n = 391), participants in a heterosexual, romantic relationship completed self-report measures of childhood experiences, romantic attachment, and IPV perpetration. Study 1 provides support for Hypothesis 1. Hypothesis 1 is replicated only for men, but not women, in Study 2. Results of Study 2 provide support for Hypothesis 2 for men and women, and Hypothesis 3 was supported for men but not women. The findings contribute to the literature addressing the association of environmental risk in childhood on adult romantic relationship outcomes.