Oil painting is a kind of mass medium with unique charm and influence, which has a wide dissemination effect. If social values can be spread through oil painting, they can have a wider and more effective effect. In this paper, the Fisher score algorithm is used to assess the degree of correlation between the color features of oil paintings and the classes they belong to. The grey scale co-production matrix is used to extract texture features of oil paintings. The regularized limit learning machine model is introduced to improve the accuracy of algorithmic feature extraction. The oil painting artworks’ feature vectors are calculated, the RELM model is employed to recognize them, and the plain Bayesian classifier confirms the accuracy of the recognition results. A simulation analysis of the oil paintings is carried out. The highest pixel value is 1880 when the luminance of the red histogram is between 10-15 in the oil painting of “Sleeping Venus”, and the lowest pixel value is 184 in this interval in the “Unknown Woman”. The four grey gradient covariance matrices eigenmeans of the oil paintings are calculated, and the eigenquantities ASM, CON, CORRLN, and IDM are rotated by 90°. The eigenmeans are 0.047, 28.766, 0.125, 0.18. Based on the results of the experimental analysis, the correlation between the characteristics of oil painting art and the values of modern society is discussed.