While physiological stress responses have been well investigated in individuals, the effects of stress on interpersonal physiological processes have been studied less frequently. In the current study, we focused on how stress affects physiological synchrony (PS) – i.e., the alignment of physiological changes across individuals – an established phenomenon characterizing social interactions. We hypothesized that PS, specifically cardiac PS, would be altered in romantic couples if one of the partners was exposed to a standardized stressor. In a preregistered study, N = 75 couples (mean age = 22.66 ±2.99, 51% female) were separated upon arrival in the laboratory. In n = 38 dyads, one partner was exposed to a laboratory stressor while the other completed a non-stressful control task (stress dyads), in the other dyads both partners underwent the control paradigm (control dyads). Afterward, partners were reunited, completing a non-verbal synchronization, a walking, and a free interaction task. Partners rated their own and each others’ affect throughout the experiment. Compared to a non-interactive baseline, PS increased during the partners’ interaction. Cardiac synchrony related to parasympathetic nervous system (PNS) activity was lower in stress compared to control dyads. Further, participants were more accurate in estimating their partner’s emotional valence if their partner was non-stressed. Our findings indicate the disruption of especially PNS-related PS. This highlights that stress is not only an intra- but also interpersonal phenomenon whose effects on the interpersonal physiology of social interactions should be further investigated.