The capacity of information processing system for human being is severely limited, but humans are proficient in searching for target information in the familiar visual scenes, in part because the task-relevant long-term memory (LTM) representations can efficiently guide attentional deployment to optimize the selection to the target and the escapement from the distractors. Hence, LTM-guided attention is key to our high level of visual performance, serving to direct our limited attentional resources efficiently. However, the issue whether irrelevant LTM representations can guide the deployment of visual attention as well as the irrelevant working memory (WM) representation is elusive yet. Therefore, we attempted to explore this issue here via three experiments. In experiment 1, participants were asked to maintain an object in LTM before the experiment initialized until to the end. During the experiment, participants were required to perform a visual search task while holding another object in WM online. In the visual search task, one of the distractor might share common features with either the representation of LTM or the representation of WM occasionally. Both the results of the response time and the first fixation proportion showed that the visual attention would bias to the distractor when sharing common features with the WM representation, displaying an classical WM-driven attentional guidance effect; however, non-guidance effect was found when the distractor shared common features with LTM representation. More importantly, the magnitude of guidance from WM representation was not affected by the simultaneously- emerged LTM representation which was regarded as a directly competitor for the attentional resources in the visual search display. In experiment 2, we manipulated the repetition times of the remembering object as the task used by Carlisle, Arita, Pardo & Woodman (2011), and aimed to test the attentional guidance from the memory representation while it was transferred from WM to LTM. The results observed an obvious attentional guidance effect from the memory representation when it was regarded as being maintained in WM (i.e., when the remembering object repeated less than three times) and this guidance effect disappeared when the memory representation was turned into LTM representation (i.e., when the remembering object repeated more than three times). In experiment 3, we required participants only keeping the LTM representation in memory system as to eliminate the possible interference from WM representation, and remain did not found any attentional guidance effect from the irrelevant LTM representation. In conclusion, the results of the present study observed a robust attentional guidance from the WM representation even when it not severing as search target template and sharing features with distractor in visual search task, in contrast, none such effect was found from the LTM representation under the same situation. These results indicated that the irrelevant LTM representation could not guide visual attention as well as irrelevant WM representation, and illustrated that the guiding process of visual attention from the representations of WM and LTM were two of distinct cognitive processes.