The mechanisms of semantic conflict and response conflict in the Stroop task have mainly been investigated in the visual modality. However, the understanding of these mechanisms in cross-modal modalities remains limited. In this electroencephalography (EEG) study, an audiovisual 2-1 mapping Stroop task was utilized to investigate whether distinct and/or common neural mechanisms underlie cross-modal semantic conflict and response conflict. The response time data showed significant effects on both cross-modal semantic and response conflicts. Interestingly, the magnitude of semantic conflict was found to be smaller in the fast response time bins than in the slow response time bins, whereas no such difference was observed for response conflict. The EEG data demonstrated that cross-modal semantic conflict specifically increased the N450 amplitude. However, cross-modal response conflict specifically enhanced theta band power and theta phase synchronization between the medial frontal cortex (MFC) and lateral prefrontal electrodes as well as between the MFC and motor electrodes. In addition, both cross-modal semantic conflict and response conflict led to a decrease in P3 amplitude. Taken together, these findings provide cross-modal evidence for domain-specific mechanism in conflict detection and suggest both domain-specific and domain-general mechanisms exist in conflict resolution.