Humans typically use three kinds of spatial reference system to understand the world: the egocentric, allocentric, and intrinsic reference frames. In the current study, we used a reconstruction task (the animals-in-a-row task) to examine the development of children’s use of the intrinsic reference frame. Japanese children aged four to six years participated in experiments to examine (1) if there is an intrinsic reference frame away from the stimulus tables, whether children prefer to use the intrinsic reference frame over the egocentric and the allocentric reference frames; and (2) whether there is an age difference in using the intrinsic reference frame in early childhood. The results revealed that Japanese children who are relative speakers used the experimenter’s position as the intrinsic reference frame, and the use of the intrinsic reference frame was related to individual preference without developmental differences.