Living in a complex and multisensory environment demands constant interaction between perception and action. In everyday life it is common to combine efficiently simultaneous signals coming from different modalities. There is evidence of a multisensory benefit in a variety of laboratory tasks (temporal judgement, reaction time tasks). It is less clear if this effect extends to ecological tasks, such as walking. Furthermore, benefits of multimodal stimulation are linked to temporal properties such as the temporal window of integration and temporal recalibration. These properties have been examined in tasks involving single, non-repeating stimulus presentations. Here we investigate the same temporal properties in the context of a rhythmic task, namely audio-tactile stimulation during walking. The effect of audio-tactile rhythmic cues on gait variability and the ability to synchronize to the cues was studied in young adults. Participants walked with rhythmic cues presented at different stimulus-onset asynchronies. We observed a multisensory benefit by comparing audio-tactile to unimodal stimulation. Moreover, both the temporal window of integration and temporal recalibration mediated the response to multimodal stimulation. In sum, rhythmic behaviours obey the same principles as temporal discrimination and detection behaviours and thus can also benefit from multimodal stimulation.