Abstract: Garment seam subjective evaluation tends to be influenced by the evaluator.For this reason,this paper used FAST system to measure mechanical properties of garment fabric at low stress,and applied main factor method to analyze the measured mechanical properties,then extracted 6 main factors as the input of neural network.Employing FCM clustering algorithm to improve RBFNN(calculating prototypes and widths of hidden nodes with clustering results),this paper proposed a fuzzy seam smoothness evaluation system.Experimental results demonstrate that the proposed approach can quickly and accurately present the evaluation result of the seam smoothness of garment made of thick worsted fabric of various structures and mechanical properties.