Abstract: This paper proposes a new automatic recognition algorithm for wool knitted clothing based on fuzzy theory. This new algorithm can recognize arbitrary clothing pieces by computer and provide evidences for continue process sheet calculation, size parts recognition and process formulation. Firstly, it does digitized description for clothing pieces by clothing pieces’ data model based on key points, and proposes a contrast method between arbitrary and standard clothing pieces. Secondly, it gives a membership function between two arbitrary and standard clothing pieces based on key points and contour points according to equal proportion transformation of clothing pieces. After normalization processing by this membership function, a general membership function for determining the classification of arbitrary clothing pieces is obtained. Finally, we compile some programs to verify above algorithm and obtain an experimental result that the recognition accuracy of arbitrary clothing pieces can reach more than 95%. From these results, we conclude that the classification of arbitrary clothing pieces can recognized more accurately by this algorithm.