Purpose – This study aimed to cluster salty snack products based on their nutritional information using the K-Means algorithm and to identify the most suitable cluster for supporting a calorie-deficit diet, addressing high obesity rates in Indonesia and consumer difficulties with nutrition labels. Methods – This study used a quantitative data mining approach, 620 products were collected from FatSecret Indonesia via purposive sampling across variables of calories, fat, protein, fiber, sugar, and sodium. The data were processed through cleaning, missing-value checking, transformation, and Z-Score normalization, followed by K-Means clustering, Elbow method optimization, and evaluation via Silhouette Coefficient and Davies-Bouldin Index (DBI). Findings – The analysis identified three optimal clusters (k=3). The overall Silhouette Coefficient was 0.431 and DBI was 1.1534. C1 (461 products) featured lower calories, fat, sugar, and sodium, making it the most suitable cluster among the identified clusters. However, only 185 of the 461 products in C1 met the defined product-level safety criteria. C2 (96 products) was high in calories and fat, and C3 (63 products) had high protein but extreme sodium levels. Research implications – The study's validity is constrained by its reliance on secondary online data without direct laboratory verification. Furthermore, the use of mean imputation for the missing fiber data may have attenuated inter-cluster variance and reduced the discriminatory role of fiber. Originality – This research provides a novel contribution in grouping salty snack products by combining six main nutritional variables along with comprehensive model performance evaluation. C1 was identified as the most suitable cluster for supporting a calorie-deficit diet.