The surface roughness and surface texture feature both are the key factors to evaluate the ground ceramics surface. The Gray Level Co-occurrence Matrix (GLCM) is introduced to extract and analyze the texture features of ground surface, in order to discover the relationship between texture features and roughness, and predict the roughness of ground ceramics according to the defined texture parameters. Four matrices about different sampling orientations are built up after the sampling offset and total gray levels are determined. By analyzing the correlation about every two features, Contrast, SS and ASM are selected to characterize the texture information of machined surface. Finally, the paper reveals the linear variation laws between Ra, Ry, S, tp and above features. The grounding quality of ceramic products also can be roughly estimated on the ground of that.