The formation of coherent multisensory percepts requires integration of stimuli across the multiple senses. Patients with schizophrenia (ScZ) often experience a loss of coherent perception and hence, they might also show dysfunctional multisensory processing. In this high-density electroencephalography study, we investigated the neural signatures of the McGurk illusion, as a phenomenon of speech-specific multisensory processing. In the McGurk illusion lip movements are paired with incongruent auditory syllables, which can induce a fused percept. In ScZ patients and healthy controls we compared neural oscillations and event-related potentials (ERPs) to congruent audiovisual speech stimuli and McGurk illusion trials, where a visual /ga/ and an auditory /pa/ was often perceived as /ka/. There were no significant group differences in illusion rates. The EEG data analysis revealed larger short latency ERPs to McGurk illusion compared with congruent trials in controls. The reversed effect pattern was found in ScZ patients, indicating an early audiovisual processing deficit. Moreover, we observed stronger suppression of medio-central alpha-band power (8–10 Hz, 550–700 ms) in response to McGurk illusion compared with control trials in the control group. Again, the reversed pattern was found in SCZ patients. Moreover, within groups, alpha-band suppression was negatively correlated with the McGurk illusion rate in ScZ patients, while the correlation tended to be positive in controls. The topography of alpha-band effects indicated an involvement of auditory and/or frontal structures. Our study suggests that short latency ERPs and long latency alpha-band oscillations reflect abnormal multisensory processing of the McGurk illusion in ScZ.