Principal component analysis (PCA) was used to explore the relationship among molar proportions of individual rumen volatile fatty acid (VFA) and milk odd- and branched-chain fatty acid (OBCFA) concentrations. Data were obtained from an experiment with 24 lactating Holstein cows examining the effects of diets containing variable proportions of forage (39, 43.5 or 48% of forage in total diet DM) and concentrate containing sunflower oil (3% in diet DM) on milk fat content and composition. The data were analyzed using the loading plot of PCA which was performed using PASW (version 18.0). Molar proportions of butyrate were included in the initial evaluation, but these data were subsequently excluded because these did not meet the necessary PCA model fit criteria. Loading plot of PCA revealed that rumen molar acetate proportions clustered with milk iso fatty acids of <15 carbon atoms, whereas rumen molar propionate proportions clustered with milk straight odd-chain fatty acids (15:0 and 17:0) and their desaturation products in the opposite side of both PCs. Concentrations of 15:0, cis-9 15:1 and cis-9 17:1 had high loadings on PC1, whereas 17:0 had high loading on PC2 which was located close to molar propionate proportions. In conclusion, rumen molar acetate and propionate proportions were closely associated with milk fat iso and straight odd-chain fatty acids concentrations, respectively, in cows fed diets containing sunflower oil.