This randomized controlled trial examined (a) the efficacy of a brief intervention designed to increase the rate of secure infant attachment, (b) the differential susceptibility hypothesis, and (c) whether maternal attachment styles moderated the expected Treatment x Irritability interaction in predicting infant attachment outcomes. Although there was no main effect of treatment, a significant Treatment x Irritability interaction revealed intervention effects for the highly irritable infants only, thus supporting one of two predictions of the differential susceptibility hypothesis: highly irritable infants would have disproportionately better outcomes than moderately irritable infants in better conditions (i.e., with intervention). When separate analyses were conducted with maternal attachment styles, we found significant three-way interactions among treatment, irritability, and each of the examined maternal attachment style dimensions (i.e., secure-fearful and dismissing-preoccupied). Specifically, with more secure mothers, beneficial effects of intervention emerged for highly irritable infants. For more dismissing mothers, the results revealed support for both predictions of the differential susceptibility hypothesis: highly irritable infants, compared to moderately irritable infants, were both more likely to be secure with intervention and less likely to be secure when in the control group. It is interesting that, for more preoccupied mothers, a treatment effect emerged only for moderately irritable infants. We discuss the implications of these findings for the differential susceptibility hypothesis as well as for early intervention.