Abstract: In order to evaluate more objectively the performance of knitting yarns, a method using modified fuzzy kernel C-Means (FKCM) clustering algorithm for processing and analyzing of knitting yarns is proposed, in which, the data of low dimension input space is mapped to high dimension feature space, FCM clustering algorithm is performed in the feature space, then the Kernel F clustering validity index is designed for seeking the fitness clustering number, and the corresponding relationship model of class sequence numbers and quality grades is constructed. By analyzing the IRIS Dataset, the result shows the modified FKCM has obtained better classification effect. When it is applied to measuring dataset, and KF index indicates it is reasonable to classify the samples into two kinds. According to the constructed relation of classes and quality grades, the quality grade of each class is acquired. The combination of modified FKCM and KF index provides an efficient data analysis method for multi-index dataset.