In this paper, a fuzzy pattern recognition technique is applied to classifying aluminium weld quality in tungsten inert gas (TIG) welding. The pattern vector includes three components, that is, the front height, the back height, and the front width of weld. Based on the values of the pattern vector, good, fair, and poor weld qualities can be automatically classified by using the fuzzy pattern recognition technique. Experimental results under different welding parameters are presented to illustrate the proposed method.