This study examined the electrophysiological correlates of the processing of the Remember/Forget cues and the successful encoding of study items in item-method directed forgetting. Subjects engaged in an old/new recognition test and an item-method directed forgetting task. Event-related potentials (ERPs) time-locked to study items and Remember/Forget cues were compared according to the subsequent recognition performance. A reliable subsequent memory effect was elicited by the study items in the old/new recognition test. In contrast, the study items in the directed forgetting task did not yield reliable subsequent memory effects. Importantly, the Remember/Forget cues gave rise to ERPs that were predictive of the subsequent recognition performance to the study items preceding the cues. The subsequent memory effect elicited by the Remember cues was more sustained than that elicited by the Forget cues and showed distinct scalp distribution during the extended period. These results suggest that study items in the directed forgetting task are maintained in short-term memory with minimal further processing until the presentation of the Remember/Forget cues. In addition, the encoding mechanisms engaged by Remember cues and Forget cues are not entirely equivalent.