To solve the problem of the uncertain style knowledge acquisition during the product designing, the information model oft style with Kansei Engineering and a reduct approach to style information based on rough sets is introduced. On the basis of the attribute analysis of form features and the image semantic extractions by the web-based style evaluation system. In the system the evaluate information of style is obtained and the decision-making table is constructed. And then the valuable knowledge rules represented the typical product style are acquired through the attribute reduction algorithm. Finally, the attribute feature sets related to the different style image could be presented. Typical Ming-style chairs as an example to show that the method is effective and feasible.