Abstract Maintenance makes machine life long more and gets no loss about a chain reaction on part broken. Repair a part that has high complex and is important. It need engineer experience to decide its maintenance process plan. For turbine blade maintenance style that has various type and a small quantity. Different broken feature have different maintenance process in the maintenance process planning stage. This thesis creates a system for this property maintenance on the process planning system on the Maintenance of turbine blades. Analyzing blade broken feature and classifying the feature to bring up a case-based reasoning and fuzzy set operation that usually see in expert systems. It uses case-based databases to save engineer experiences and analysis to decide critical maintenance process factors. The critical factors can help find out similarity case and modulate a newer case. It needs to create a proto-type system for these purposes. The first, classifying turbine blades broken feature and coding the repair process for different maintenances. The second, using fuzzy set language to determine case¡¦s similarity and finding out algorithm. Integrating product data management system and the system. The research uses a company as case study to test and verify the proto-type system.