Abstract: Machining process knowledge excavation and reuse is difficult because process knowledge is hidden in the geometry variation of intermediate process model sequence. A generation method of geometry variation sequence of intermediate process model, a process similarity measure algorithm, and a feature matching algorithm were presented in this paper. The geometry variation between the former procedure and the latter procedure is extracted and expressed, and the geometry variation sequence is set based on the precedence constraints of processing sequence. An algorithm of measuring similarity between geometry variation sequences is developed to evaluate the process similarity of parts. The geometry variation elements which is correlated with the formation of machining feature is extracted from the geometry variation sequence, and a new machining feature model is established. By applying feature model matching algorithm, features recognition and matching which integrates the formation process of features is realized. Experimental result shows that the effectiveness of proposed method is verified.