Food packaging design plays a crucial role not only as a protective medium but also in shaping consumers’ emotional perceptions. However, many existing studies still focus primarily on aesthetic aspects without providing a systematic analytical framework to translate affective responses into design parameters. This study aims to develop a visual packaging design concept for cheese sauce by integrating Kansei Engineering with data-driven analytical approaches. Theoretically, this study proposes an exploratory framework that links perception weighting (TF-IDF), dimensionality reduction (PCA), and design pattern modeling (Association Rule Mining/ARM) within a unified analytical process. Kansei Words were collected from respondents (n ≥ 30), analyzed using TF-IDF, and reduced using PCA to identify dominant perceptual orientations. Packaging design elements were examined through morphological analysis of 25 product samples and formulated using ARM. The results indicate that consumer perception is strongly dominated by a Premium design orientation, explaining 75.16% of the variance, while the Eco-friendly orientation emerges as a secondary dimension with limited contribution. The association patterns suggest that non-transparent packaging, full label coverage, and top brand placement represent the most consistent configuration within the dataset. The findings suggest that integrating Kansei Engineering with data-driven methods can serve as a conceptual basis for formulating packaging design directions. However, this study remains exploratory, with limitations related to sample size, data homogeneity, and the absence of preference validation for the proposed design, indicating the need for further empirical testing.