Sunflower seeds are rich in oil and oleic acid, thus having valuable nutritional properties. Sunflower is also resistant to dry conditions and can adapt easily to harsh environmental conditions. Physico-chemical properties play a great role in classification, grading, and quality assessment of sunflower seeds. In present study, six different machine learning algorithms (decision tree, DT; random forest, RF; support vector machine, SVM; multiple linear regression, MLR; Naïve Bayes, NB; and multilayer perceptron, MLP) were used to evaluate the classification performance for six different sunflower oilseed varieties. Additionally, characteristic properties of the oilseeds were evaluated by multivariate tests (MANOVA) and discriminant analysis. The best accuracy values were recorded as 80.16, 79.68, and 78.89 for RF, SVM, and MLP, respectively, and the lowest MAE value (0.088) was observed in NB. The MANOVA, Wilks’ lambda, and Pillai Trace statistics revealed that differences in physical attributes of the sunflower varieties were significant (p<0.01). Colombi and Transol varieties with the lowest Mahalanobis distances had the similar attributes.