Temporal dimension is very important in text comprehension. And a number of previous studies demonstrated readers tracked time information during text reading. Two hypotheses elaborate how readers deal with the flashback information in text comprehension. The chronological hypothesis indicates that readers integrate the flashback information into their representation at its chronologically appropriate position, while the background hypothesis considers that readers interpret a flashback as providing background information for the situation described before. Claus and Kelter (2006) testified both the chronological hypothesis and the background hypothesis using a probe-recognition task paradigm. Participants read short flashback texts which included four events (E1, E2, E3, E4) and the events were described in a different order (E2, E3, E1, E4). The duration of E2 was manipulated, and at the end of each passage readers was tested whether they accessed E1. The results found that E1 was much more accessible when it occurred in a relatively short time than longer, which supported the chronological hypothesis. However, the specialty of experiment materials may contribute to above results. Firstly, the temporal distance of reading text was less than a day. Leng et al (2004) found temporal distance in narrative influenced construction and updating of reader’s situational model. If enlarged temporal distance in narrative containing flashbacks, readers may choose different strategy to construct and update situational model. Secondly, there were obvious grammar markers in reading materials. The significant grammar markers for tense would give readers big time clue, which supported the chronological hypothesis. Two experiments were conducted to investigate how readers represented flashback texts with different time information in Chinese, which did not have grammar markers for tense. Experiment 1 explored how readers represented flashback information when the text was about events happened in a few hours. The result indicated that E1 was less accessed in a long duration condition compared with the short duration condition. It suggested that readers represented the text information and supported the chronological hypothesis. Experiment 2 explored how readers represented flashback information in narrative texts which described events happened in a few days, and the result indicated that there was no significant E1 accessibility difference between the long duration condition and the short duration condition. It suggested that readers took flashbacks as the background of narrative, which supported the background hypothesis. These two experiments suggested that the time information of text influenced how readers represented flashback information.