Visuo-tactile cross-modal congruence effects (CCEs) refer to impaired tactile processing when task-irrelevant visual cues are presented close to the body but in a spatially incongruent location. This interference extends to visual cues appearing far from the body and on distant objects when individuals are reaching toward those objects or observing someone else grasping them. These effects highlight the impact of motor processes on cross-modal integration. However, it remains unclear whether real-time interpersonal interactions, which require continuous prediction of others’ actions, influence the spatial boundaries of CCE. Across three behavioral experiments, we found that interpersonal motor interactions modulate cross-modal integration. Specifically, action-related visuo-tactile CCE emerged only when the location of the visual cue was incongruent with the predicted movement of an interaction partner. These results suggest that interpersonal sensorimotor predictions may shape basic spatial mappings of cross-modal interference, indicating that intersensory processing is dynamically influenced by social interaction demands.