Despite the growing use of k-means clustering in cattle breeding research, its application to weaning performance traits in Limousin young bulls remains limited. This study evaluated the effectiveness of k-means clustering for identifying distinct groups based on age, live weight and morphometric records in a Hungarian Limousin nucleus herd. A total of 146 bulls weaned at an average age of 213 ± 2.03 days and live weight of 263 ± 3.31 kg (mean ± S.E.M.) were analysed. Age, live weight, withers height, rump height, back length, shoulder width, hip width and pin width were evaluated using k-means clustering. The optimal number of clusters was determined using the Elbow method, Average Silhouette Width analysis, the Calinski–Harabasz index, and the Davies–Bouldin index. Three clusters were identified, representing low (213 kg), intermediate (260 kg), and high (309 kg) body size groups. Live weight showed the strongest contribution to clustering (F = 387.3), followed by rump height (F = 85.6) and withers height (F = 82.8). Back length (F = 13.7) and pin width (F = 12.6) contributed the least. The greatest separation was observed between Cluster 1 and Cluster 3 (Euclidean distance = 99.26). Cluster 3 showed the highest within-cluster homogeneity, with no outliers and the lowest dispersion (SD = 10.67). These findings support data-driven phenotypic preselection strategies; however, future research should incorporate additional herds, economically relevant traits, and genomic data to enhance generalisability.