Background and methods: In two birth cohort studies with genetic, sensitive parenting, and attachment data of more than 1,000 infants in total, we tested main and interaction effects of candidate genes involved in the dopamine, serotonin, and oxytocin systems (DRD4, DRD2, COMT, 5-HTT, OXTR) on attachment security and disorganization. Parenting was assessed using observational rating scales for parental sensitivity (Ainsworth, Bell, & Stayton, 1974), and infant attachment was assessed with the Strange Situation Procedure. Results: We found no consistent additive genetic associations for attachment security and attachment disorganization. However, specific tests revealed evidence for a codominant risk model for COMT Val158Met, consistent across both samples. Children with the Val/Met genotype showed higher disorganization scores (combined effect size d = .22, CI = .10-.34, p < .001). Gene-by-environment interaction effects were not replicable across the two samples. Conclusions: This unexpected finding might be explained by a broader range of plasticity in heterozygotes, which may increase susceptibility to environmental influences or to dysregulation of emotional arousal. This study is unique in combining the two largest attachment cohorts with molecular genetic and observed rearing environment data to date.