Background Risk for posttraumatic stress disorder (PTSD) is thought to be mediated by gene × environment (G × E) interactions that affect core cognitive processes such as fear learning. The catechol‐O‐methyltransferase (COMT) val158met polymorphism has been associated with risk for PTSD and impaired fear inhibition. We used a large, relatively homogenous population to (1) replicate previous findings of poor fear inhibition in COMT Met/Met carriers with PTSD; (2) determine if COMT association with fear inhibition is moderated by childhood trauma (CT), an environmental risk factor for PTSD; and (3) determine if COMT is associated with altered fear processes after recent exposure to combat trauma. Methods Male Marines and Navy Corpsmen of European‐American ancestry were assessed prior to (n = 714) and 4–6 months after deployment to Afghanistan (n = 452). Acquisition and extinction of fear‐potentiated startle, childhood and combat trauma history, and PTSD diagnosis were assessed at both time points. Results Before deployment, Met/Met genotype was associated with fear inhibition deficits in participants with current PTSD; however, this association was dependent on CT exposure. After deployment, combat trauma was associated with a modest reduction in fear extinction in Met/Met compared with Val/Val carriers. There were no associations of COMT genotype with fear extinction within healthy and non‐traumatized individuals. Conclusions These findings support the hypothesis that G × E interactions underlie associations of COMT val158met with fear inhibition deficits. These studies confirm that Met/Met carriers with PTSD have poor fear inhibition, and support further research in understanding how this polymorphism might impact response to extinction‐based therapies.