This study aimed to determine which nutritional traits are a key to indicate the nutritive value of some sorghum cultivars by chemometric techniques such as principal component analysis (PCA) and cluster analysis (CA). Based on their genotype, seven sorghum (Sorghum bicolor L. Moench) cultivars (sorghum [Rox and Early Sumac], sudangrass [Gözde] and sorghum × sudangrass hybrid [Jumbo, Grazer, Hayday and El Rey]) were denoted. Nutritional traits used were comprised of neutral detergent fibre (NDF), acid detergent fibre (ADF), crude protein (CP), Ca, P, K and Mg that determined by near-infrared spectroscopy analysis and digestible dry matter (DDM), dry matter intake (DMI), metabolizable energy (ME) and relative feed value (RFV) calculated by empirical equations. The NDF and ADF contents showed a negative relationship with ME and RFV. The CP content was negatively correlated with NDF and ADF but positively correlated with ME and RFV. The PCA generated two significant principal components (PCs). PC1 (59.51%) and PC2 (20.31%) described 79.83% of the total variation with eigenvalues of 6.55 and 2.23 in the sorghum cultivars, respectively. The PC1 was more representatives of the cultivars (Grazer, Hayday, Early Sumac and Gözde) and the calculated traits (DDM, DMI, ME and RFV). PC2 was characterized by Ca and Mg, while ADF, NDF and K with negative loadings in each of the PCs were the most representatives of most cultivars. CA grouped cultivars and traits into two clusters. The P, DDM, DMI, ME and RFV were the most important among components, defining important traits of cultivars to improve feeding of ruminants.