This study aimed to objectively select the most appropriate clustering method and number of clusters for describing dietary patterns using a stability-based methodology. Cluster analysis was performed on dietary data from the NESCAV study using K-means and principal component analysis (PCA). The most stable solution was obtained using K-means clustering with three clusters. This identified three main dietary patterns in the population: a "Convenient" and "Non-Prudent" pattern associated with higher cardiovascular risk, and a "Prudent" pattern associated with decreased risk.