We apply a preference rule extraction framework to a makeup support system that utilizes interactive evolutionary computation to better understand users’ makeup preferences. The proposed system derives preference rules in the form of fuzzy rules that represent users’ perceptions using descriptive adjectives. Compared with our previous system, the proposed system enhances direct manipulation capabilities, allowing users to adjust the darkness of specific cosmetic features to refine their makeup design. We conducted an evaluation experiment with real participants, measuring the time required to use the proposed system, the satisfaction levels of the generated makeup patterns and preference rules, and the similarity between each generated preference rule. The results indicate that although the proposed system requires more time for direct manipulation compared to the conventional system, it achieves the higher satisfaction levels for both generated makeup patterns and preference rules.