Abstract：Though cognitive advantage for self-related information has been well established in a number of studies using various stimuli. However, the mechanism for this effect remains unclear. In order to understand the function of attention in cognitive advantage for self-information, this study used the attention network test (ANT) to evaluate the difference between self-related information with friends-related information from three terms about the attention function including alerting, orienting and executive control. In experiment 1, all participants completed a version that presented targer and flanker stimuli, the target array consisted of a central target stimulus and four flanker stimuli. Each target was preceded by one of four cue conditions: a center cue, a double cue, a spatial cue, or no cue. Participants were instructed to pay attention to the face in the middle and press whichever button matched the self-or friend-face. Results showed that participants had a stable advantage in processing self-face, and the efficiency of orienting on self-face was significantly larger than friend-face. In experiment 2, we used recent developed self-associated learning approach to exclude the interference of familiarity. The result was consistent with experiment 1, the self-referential shapes had advantages on the efficiency of orienting network. But no difference was found in the process of the other two shapes (friend-referential shapes and nonsocial-meaning shapes). It implied that the improvement of orienting network on self-information processing was not because of familiarity but the important meaning contained in self-information. In experiment 3, We found that when the task was to determine the color of the face, the processing priority on orienting network of self-face was disappeared. It indicated that the task requirement influenced the cognitive advantage for self-information on orienting network. The major findings of this study were summarized as the following three aspects: Firstly, in the three attention networks, orienting network showed the processing priority of self-information, and therefore played an important role in self-processing advantage. There was no special bias of alerting network or executive control network in self-information processing. Secondly, the task-directed orienting network influenced the presence or absence of self-processing advantage.