Research on text comprehension is an important part of cognitive psychology, Situation model is considered as the higher level of text representation, so it attracts the focuses of many researchers. The spatial dimension of situation model has been explored most often. Layout-learning & Anaphora resolution is an important way to explore the representation of spatial distance in text comprehension. Our paper explored the rules of distance representation in three experiments and advanced some hypotheses. In Experiment 1, we used the method of layout-learning & Anaphora resolution while controlling the path distance (including category distance and measure distance) and linear distance, with the number of turns as the independent variables, to examine whether the route-angularity effect occur. During the Experiment1, those who are experimented first should study the layout diagram of the company, and memorize the location of each room and the specific items in each room until they feel they can describe the location of each room and the specific items within the room in detail. During the Experiment1, the observers were first asked to learn the layout diagram of a fictional company, including the specific locations and names of each room. Then a reading comprehensive test was conducted to make sure the observers could remember these details. In Experiment2, we used a pure natural discourse reading test, which was traditionally supposed to construct the spatial situation mode, to determine whether there is still route-angularity effect in discourse comprehension under such a mode, and the whether the effect, if any, was modulated by the complex levels of spatial in description. In the natural discourse reading test, we controlled the path distance and linear distance, with the number of turns as the independent variable, and require the observers first read some discourses relevant to route-angularity, and then finish the reading comprehension questions directly. Experiment 3 used the mode that the leading character walks synchronous dynamic with the discourse, with the help of eye detecting, forcing the observers follow the leading character’s route. Controlling the number of turnings and route distance, with the linear distance between the current location and the anaphora position as independent variable, we aimed to probe the observers’ psychological mechanism when they construct spatial situational mode in complex spatial description. The observers were showed dynamic route map while reading the discourses, and after that, they have to judge the mark in the map. The observers’ reading time of the target sentences, the total fixation time in anaphora words, and reaction time of judging the mark were recorded. A hundred and ten college students with normal vision participated, with sixty in Experiment 1 and Experiment 2, with fifty in Experiment 3, respectively. In Experiment 1, we used the method of layout-learning & Anaphora resolution which was explained in the foregoing paragraph. In experiment 2, participants read narratives describing location of a small town, and in each site there was an object. By motion sentences, the participants' attentions were removed from one site to another. Reading times of the following target sentences that refer to objects in one of these site by means of a definite noun phrase can imply whether participants established a spatial representation of text or not. In experiment 3, using eye movement monitoring, display the text and the routes synchronous, forcing the subjects to follow the protagonist in real time, Separated category distance and metric distance and the number of angularity, the results showed that when the category distance and metric distance and the number of angularity were all the same, the Straight-line distance influenced the time of anaphora resolution. All the data were analyzed with repeated-measure ANOVA. In Experiment 1 and Experiment 2, the general trend is that the observers spent more time on reading the target sentences as the increasing appearances of route-angularity when the discourse spatial description is simple (angularity routes<2). By contrary, when the discourse spatial description is comparatively complicated (angularity routes >2), there would be no any trace of the route-angularity effect but instead the reversal effect would appear. In Experiment 3, the Fixation time of the anaphora words was longer when the straight stretch was farther in the complicated spatial condition, and also the time spent on reading the target sentences and the judging time. Integrated the three experiments, we can prove that route-angularity does effects on distance representation in text comprehension. When the description of location is simple, the spatial situation models which readers constructed were route-chart type, they completed anaphora resolution through reverse searching; when the description is more complex, readers prefer to construct a map-based space situation model, and find the shortest straight line distance through the space bridge.