Integration of information from multiple senses is fundamental to perception and cognition, but the neural activity of multimodal audiovisual integration remains unclear. This study used event-related potentials (ERPs) to demonstrate that onset synchronous task-irrelevant auditory stimuli affect the audiovisual integration. The behavioral results showed that the responses to audiovisual target stimuli were faster than that to unimodal visual target stimuli. Moreover, the ERPs were recorded in response to unimodal auditory (A), unimodal visual (V) and bimodal (AV) stimuli. Cross-modal interactions were estimated using the additive [AV - (A + V)] model. Four ERP components related to audiovisual integration were observed: (1) over central and occipital areas at around100 to 160ms; (2) over the central and occipital areas at around 160 to 200ms; (3) over the occipital areas at around 200 to 240ms. (4) over frontal-central areas at around 280 to 320ms. These findings confirmed the main neural activity of audiovisual integration. In addition, our study provided evidence that multimodal integration can be generated even if the auditory stimulus was task- irrelevant.