Multivariate statistical techniques were used to differentiate two popularly consumed grape cultivars in Taiwan, Kyoho (Vitis labruscana) and American wild bunch (Vitis aestivalis), based on their phenolic content and antioxidant properties. Total phenol and flavonoid contents, and polyphenolic profile were estimated by the Folin-Ciocalteu, colorimetric method, and ultra performance liquid chromatography, respectively. Antioxidant activities were determined by 2,2 diphenyl-1-picrylhydrazyl, 2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid) scavenging activity, ferric reducing antioxidant power (FRAP), nitric oxide scavenging activity (NOSA), hydrogen peroxide scavenging activity, and total antioxidant activity methods. Data were processed using multivariate analysis of variance, discriminant analysis (DA), principal component analysis (PCA), and cluster analysis. Variations in polyphenolic compositions were attributed to higher TFC (3.42 mg QE/g), ABTS • (79.08%), NOSA (27.62%), total antioxidant (77.91%), and superoxide (46.55%) radical scavenging activities of Kyoho grape extracts; and higher inhibition capacity by DPPH • (97.26%), FRAP (4.66 mM FeSO 4 /g), and hydrogen peroxide (40.52%) scavenging methods for American wild bunch grape extracts. Two discriminant functions in DA reported the contribution of each phenolic and antioxidant activity in discriminating the cultivars. The results of PCA analysis reported 91.35% of the total variance, which was a suitable approach for evaluating similarities between both cultivars. Statistical analyses confirmed Kyoho and American wild bunch grape extracts differed from each other in terms of polyphenolic composition and antioxidant activities. Therefore, the application of multivariate statistical techniques used in the study provided further understanding on assessment of polyphenols and antioxidant activities of these grape cultivars.