This study examined the effect of harmonization between word meaning and typography impression on implicit memory. In a preliminary experiment, we asked undergraduate students the degree of harmonization (“how much does the typography match its word meaning?”) for 153 words, showing them two lists (hiragana and kanji) with three different typographies. Based on the results, an incidental memory experiment was executed with three experimental factors: the scripts at learning (hiragana, kanji, or no learning), with/without the harmonization at learning, and the typography at testing (same as learning, different neutral typography), as within participant factors. Both a word completion task and an old/new recognition test were executed using only hiragana. The result showed that harmonization facilitated implicit memory with hiragana-learned words, but implicit memory was reduced for kanji-learned words. Morphologically-based processing occurred when word meaning and typography were harmonized. Conceptual implication of this harmonization effect will be discussed.