Abstract: From the outside,the machining process seems a process sequence formed by processing operations one by one.But from the inside,it is a geometry variation process from the initial blank model to final CAD model.It is clear that individually measuring the similarity of processing operation sequences or geometry variation sequences will reduce the effectiveness of machining process similarity retrieval.To solve this prob-lem,a multidimensional similarity measurement method of machining processes driven by sign entropy is pre-sented.First,a similarity measurement way of sequence based on sign entropy is proposed,which can directly realize the similarity measurement of processing operation sequences.Then,the geometry variation sequence corresponding to processing operation sequence is established,and further achieved the goal of measuring the similarity of geometry variation sequences in the support of the sign entropy way.Finally,make the above two similarity measurement results are fused together,and thus evaluating the similarity of machining processes is evaluated multidimensionally and accurately.Case studies show that the evaluation result by using proposed method to measure the similarity of different kinds of parts is closer to actual manufacturing process and cog-nitive experience,which proves that the proposed method is feasible.