A new approach of parting design of turbine blade based on feature decomposition and parting template library is put forward. The exceeding obstacles of turbine blade with complicated outer structure is addressed. First

automated feature recognition and man-machine interaction feature recognition are adopted to extract decomposition features of cavity. The core block of turbine blade is extracted to simple features. Second

for each simple feature

the parting design approach is selected automatically from parting template library. This new approach enhances the design effi ciency of turbine blade mould and sets a foundation of automated design of precise casting mould of turbine blade.