One of the assumptions of intergenerational family therapy is that how a person thinks and talks about family-of-origin experiences has important implications for current family relationships. The Adult Attachment Interview (AAI) is a measure of attachment security based on how coherently the person can discuss attachment experiences in childhood. This study examined the relationship between attachment security, as measured by the AAI, and couple interaction, as measured by the Georgia Marriage Q-Sort (GMQ), in a sample of 28 couples in therapy for relationship problems. During a conflict resolution discussion, those individuals who were less coherent in discussing their family of origin expressed more negative affect, less respect, less openness, more avoidance, and less willingness to negotiate when interacting with their partner. No evidence of intra-couple effects or "buffering" was found. The findings support a key assumption of intergenerational approaches to family therapy and suggest that applying attachment theory is a promising direction for refining and developing new interventions for couples.