Item memory and source memory are two important aspects of episodic memory. Source memory is defined as the time, place and background information of the object that we should remember. In event-related potentials (ERPs) researches, when participants were asked to retrieve additional information about a studied stimulus in a source memory test, there are two kinds of old/new effects: early old/new effect and late old/new effect. The early old/new effect which appears across widespread regions of the scalp, but maximal at centroparietal scalp sites is observed from around 200 to 400 ms after stimulus onset. The later old/new effect over prefrontal cortex (PFC) begins around 700 to 800 ms after stimulus onset. It is well established that source memory retrieval engages later old/new effect more than simple item memory. Some researchers think that the early old/new effect which depend on the integrity of the medial temporal lobe and diencephalon indexes successful item recognition, and the late old/new effect reflects the executive process. Thus，this late-onset PFC old/new effect play an important role in source retrieval. So far, it is still not clear why source retrieval places greater demands on prefrontal cortex than old/new recognition judgements. In the current experiment, using event-related potentials technique, we explored the effects of encoding task on source memory and its neural correlates. Seventeen college students (7 male, 10 females) took part in our experiment. Stimuli were 640 digitized pictures in which an object was superimposed on a background. During the study phase, subjects were received two kinds of tasks. For one kind of task, they were asked to make judgments about the adaptability between items and backgrounds. For the other, they were received a size-judgment task, i.e., whether the items (object) were larger or smaller than the CRT monitor. In the subsequent test phase, there were four categories stimuli (“old/ old”, “old/ new”, “new/old”, “new/ new”). Subjects were told to determine whether the picture was “completely old” “part old” or “completely new”. The ERPs for “completely old” and “completely new” were averaged separately. We supposed that the early old/new effect were greater for congruity-judgment task than size-judgment task, and the prefrontal old/new effect was insensitive to encoding task and manipulation of difficulty. The behavioral results showed that, in the study phase, the reaction times of congruity-judgment task were slower than size-judgment task. In the test phase, the source accuracy of congruity-judgment task was reliably better than that of size-judgment task for the “completely old” and “part old” conditions while the reaction times of congruity-judgment task for “completely old” condition were significantly faster than that of size-judgment task. The ERPs results showed that early and late old/new effect were observed under both encoding conditions. It was found that during about 600~700ms after stimulus onset, the amplitude of ERPs of congruity-judgment task was more positive than that of size-judgment task at FPz, and the same pattern also appeared from 700~800ms, at Pz and P4. Based the above results, conclusions could be drawn that (1) early ERP old/new effect and late prefrontal ERP old/new effect may index two important processes in source memory retrieval; (2) the influence of encoding task on source memory was mainly manifested at the more posterior brain cortex; (3) the late prefrontal old/new effect did not co-vary with the retrieval difficulty. This research promoted our understanding the neural mechanism of retrieval process of episodic memory.