It is urgent to solve the problems of low efficiency, high cost and lack of system integrity when PFMEA identifies potential process failure modes in the process of multi-variety and small batch production. Current study innovatively proposes a machine identification method of potential process failure modes based on the combination of process constituent elements (PCE) model and feature case-basedreasoning represented by extension. Firstly, the specific description of the process content is developed with the PCE and the way of expression is standardized through the method of data mining. Then, the PCE of the normative knowledge representation of extension feature cases were constructed, and natural language processing (NLP) technology is employed to solve the feature similarity between the same PCE after the specification of the sentence and chunk features respectively, and case-based reasoning and extension operator are applied to carry out analogical reasoning and calculation for the feature cases with high similarity degree of feature attributes, so as to realize the machine identification of the potential process failure modes in the manufacturing process. Finally, the proposed method is specifically applied to the three parts assembly process of an aircraft assembly to verify the effectiveness and applicability of the proposed method.