Abstract: To study the primary components and the corresponding attributes of the tactile texture space of textile fabrics, this paper, based on the semantic experiments of implicit memory, obtained tactile texture descriptors and examined the relations between the frequency and average order of the descriptors. On this basis, the method of cluster and factor analysis in data mining is used to establish the perceptual dimensions of tactile textures from the perspective of cognitive science. The results from cluster analysis show that roughness-smoothness, hardness-softness and thinness-thickness are three dimensions of the tactile texture space. The fourth perceptual attribute would possibly be related to the viscidity. The results from factor analysis indicate that the cumulative contribution ratio of the five sets of factors reaches to about 67.69%, indicating that the results of the cluster analysis can explain the majority perceptual attributes of the tactile texture of fabrics.