Abstract: Owing to the fact that uniform evaluation criterion of the neckline performance of woolen sweaters is not available presently,the performance of necklines of woolen sweaters is classified objectively by comfortableness and durability through testing the stretchability of the neckline.The membership function is constructed,and the optimal fuzzy division could be gained by iterative algorithm,then the samples are classified on the principle of the maximum membership grade,and finally the classification rationality and cluster validity are verified by FE index and the maximum value of fuzzy division entropy among kinds.By application of the cluster method based on fuzzy division,the effect of subjective factors is alleviated,and the effectiveness of data processing is improved.The classified samples could be the training samples for discriminate functions of new samples for evaluation system.