In this study, discrimination between different art categories was presented. To be able to classify different art images, features capable of including the characteristic properties of art types were extracted. Extracted features are based on RGB histogram characteristics, coarseness and edge ratio in the images. Obtained features were used in different classifier structures and an instance based learning algorithm (K-Nearest Neighbor) was preferred to be used in the classification step due to its high and robust classification performance. Obtained results show that the extracted features are highly capable of representing the characteristics of the arts.