The effect of temporal order structure and cog nitive content variables on understanding the instructional sentences of ordering syllogisms were studied in three experiments. In each experiment, 5 and 6 years old children participated in a task to solve 32 spatial ordering linear syllogisms containing the relation words: “in front of” and “behind”. The present author thought that the task-structure of syllogisms was composed of such two factors, as the type of relational inference and the direction of relational operation, and used those tasks as experimental tasks. Two types of sentences that had almost the same meaning but different grammatical structure, especially in temporal order position of relation words, were used. The main findings were as follows. (1) Such a sentence that the subject of the second prose was the logical actor, or that relational operation of both proses were based on the same anchor point but were different in direction, were understood easily (in Exp. I). (2) In spite of having grammatical complexity, such a type of sentence with a relation word appearing early was easily understood (in Exp. II). (3) It was difficult, for Ss who had low ability to grasp the relatedness of visual image, to understand the syllogisms task containing relational operations based on the different anchor point (in Exp. III). These findings suggested that comprehensibility of these relational statements did not depend on the grammatical complexity of the sentence but on the content of sentence preparing cognitive framework that made Ss easily create a whole image of arrangements. Finally, from these findings and suggestions, it was proposed that the processes of understanding these relational statements consisted of the following subprocesses: on hearing the first prose Ss create a spatial image based on relational operations, then when Ss hear the second prose, they use the spatial image as a cognitive framework to create a whole image. Based on these two processes, Ss understand the meaning of the whole sentence.