Abstract: Due to limited fault characterization method, there is no effective method to comprehensively evaluate fault sealing in complex fault-block oil and gas fields. The volume-based structural modeling technology can accurately establish a 3D geological model without being limited by the quality of seismic interpretation. According to the 3D geological model, the fault sealing characteristics are described on the fault plane from five aspects (geometry, asperity, fault offset, juxtaposition and fault clay prediction). Then a neural net algorithm is used to comprehensively classify and evaluate various indicators of fault sealing. Through the application of samples, it provides ideas for high-precision fault sealing analysis in the development stage. It shows that fault sealing analysis of complex faults based on volume-based structural modeling has obvious advantages over traditional methods.