A considerable number of studies on mental image have been reported from various fields. To the author’s best knowledge, however, no model of mental image has been proposed from the viewpoint of computer science. Mental image directed semantic theory has proposed a mental image model as a quasi-symbolic image system to be formalized in predicate logic. In mental image directed semantic theory, word meanings are treated in association with mental images, not limited to visual but omnisensory, modeled as so-called loci in Attribute Spaces which have extremely abstract quasi-symbolic images. From the viewpoint of artifact, an attribute space stands for a certain measuring instrument or sensor, like a chronograph, barometer or thermometer, and the loci represent the movements of its indicator. Mental image directed semantic theory has employed a set of temporal conjunctions denoting both logical conjunction and temporal relations between loci by themselves, because these must be considered simultaneously in locus articulation.