Audiovisual associative memory and audiovisual integration involve common behavioral processing components and significantly overlap in their neural mechanisms. This suggests that training on audiovisual associative memory may have the potential to improve audiovisual integration. The current study tested this hypothesis by applying a 2 (group: audiovisual training group, unimodal control group) × 2 (time: pretest, posttest) design. The audiovisual training group received an adaptive audiovisual associative memory training protocol, whereas the active control group undertook an adaptive unimodal item memory training protocol. The training duration for both protocols was seven days, with one and a half hours per day. Before and after training, both groups underwent testing for audiovisual integration through event-related potential (ERP) measures in a detection task that required rapid detection of auditory, visual, or audiovisual stimuli. Behavioral results revealed a training-general effect, as both groups exhibited faster response times in unimodal stimulus detection after 7-day training. EEG results revealed a consistent pattern in which the amplitudes of both visual and auditory evoked components decreased following the training. Moreover, a training-specific effect was found for sensory integration, with the effect of audiovisual integration being modulated in the earlier stage of processing for the audiovisual training group. Such a modulation was not found in the unimodal control group. These findings indicate that training on higher cognitive functions could penetrate to both unisensory processing and the integration of sensory information, which may potentially aid in the development of more targeted and efficient interventions for perceptual processes.