The present study focuses on the comprehension and production of non-literal comparisons (NLC) via visual means in three age groups: kindergarten, second grade and fifth grade students. Although non-literality is a cognitive process, the educational system does not take advantage of it in pedagogy, especially before the fourth grade. The research tools designed from scratch were based on the three categories of perceptual, visual similarities--features: Static-specific features (shape, size, color), static-general features (e.g., age: young-old) and dynamic features (e.g., motion). The results revealed that all of the children were capable of comprehending and producing NLC via visual stimuli, despite the assumption that non-literality is mastered only by older children. Furthermore, it appears that the visual aids play a key factor in both comprehension and production of NLC. Both comprehension and production of NLC increase with age, but production increase occurs at a very slow rate. Despite the fact that two separate studies were executed by the same researchers, there are slight indications that lead us to conclude that comprehending NLC is simpler than producing them. Another result is that NLC based on the static-specific features (shape, size and color) were not easier than the static-general features and dynamic features to comprehend and produce, as it has been supposed by previous studies on children's language development and thought (e.g., Clark, 1993; Winner et al., 1988).