Prospective memory (PM) refers to the task of remembering to perform an intended action at some point in the future. PM is divided into event-based prospective memory (EBPM) and time-based prospective memory. However, there are other forms of complicated PM in real life, one of which is that intended action is based on target event (i.e., PM target) and the time of PM targets occurrence is known. This form of PM is called mixed time-based and event-based prospective memory (MPM). There is no research on the neural correlates of this form of MPM. This study aimed to examine brain activity related to MPM with event-related potentials (ERPs) technique. Fifteen junior undergraduates (eight female, seven male) aged 18-22 years (M=19.7) participated as paid volunteers. A typical paradigm for studying prospective interference effect was used in this experiment. Three conditions were tested. In the baseline condition subjects had no PM task and merely performed 1-back task. In one PM condition subjects were told when the PM target would occur (named MPM condition), whereas in the other PM condition, when the PM target would occur was unknown (named EBPM condition). The reaction time of PM in the MPM condition was significantly faster than that in EBPM condition, and the accuracy of PM in the MPM condition was significantly higher than that in EBPM condition. Reaction time and the neural correlates of 1-back task of the three conditions were also analyzed. As subjects were told that PM targets appeared after 10 minutes in the MPM condition and they could estimated that the target time interval was in the third block. To examine whether subjects’ processing strategy and the neural correlates changed before and after entering the third block, all tasks were examined by two time intervals. The first and second block was named non-target time interval and the third block was named target time interval. Results showed that the reaction time of 1-back task in baseline was significantly faster than that in all blocks of the EBPM condition but only faster than that in the target time of the MPM condition, which means preparatory attentional process was engaged in PM target persistently in all blocks of the EBPM condition, whereas selectively engaged only in the target time interval in the MPM condition. No prospective interference effect existed in the non-target time interval in the MPM condition. ERP data showed that the waveforms of 150-400ms after stimulates occurred were significantly smaller for the baseline than the EBPM in all blocks and MPM condition in the target time interval. The topographical maps suggested that the components reflected activation mainly in the frontal lobe. The waveforms of 200-300ms of MPM condition in the non-target time interval were greater than that in the baseline, and the topographical maps also suggested that the components reflected activation mainly in the frontal lobe. The results might suggest that MPM was in a high activation state in brain when it was out of working memory. The results indicated that frontal lobe was engaged in monitoring PM targets persistently in EBPM condition, and selectively engaged in the target time interval in the MPM condition.