Alcohol use disorder is highly prevalent and high risk of relapse remains a significant treatment challenge. Therefore, the utility of human laboratory models of relapse to further the understanding of psychobiological mechanisms that precipitate relapse risk and allow testing of novel interventions could be of benefit in expediting the development of effective treatments to target high relapse risk. Stress is a risk factor for the development of AUD and for relapse, and furthermore, chronic alcohol use leads to adaptations in central and peripheral stress biology. Here, we review our efforts to assess the integrity of these stress pathways in individuals with alcohol use disorder and whether adaptations in these systems play a role in relapse risk. Using validated human laboratory procedures to model two of the most common situations that contribute to relapse risk, namely stress and alcohol cues, we review how such models in the laboratory can predict subsequent relapse, and how we can measure specific identified biobehavioral markers of relapse effectively and ecologically in the real world. Finally, we discuss the significant implications of these findings for the development of novel and effective interventions that target stress dysregulation and craving as risk factors to treatment.