Seismic structure interpretation is the compulsory step for 3D seismic structure modeling, stratigraphic features analysis, and 3D reservoir modeling. The modern 3D seismic surveys usually cover up to hundreds of square kilometers with thousands of inline and crossline vertical slices. Manual seismic structure interpretation (horizon and fault interpretations) on thousands of inline and crossline vertical slices is a time-consuming and tedious task. My dissertation focuses on developing new algorithms and workflows to automatically extract horizon surfaces and fault surfaces from the 3D seismic data. Most automatic horizon extraction algorithms are based on the seismic reflector dip attribute. The quality of extracted horizons is highly affected by the accuracy of the seismic reflector dip. However, the seismic reflector dip attribute is usually inaccurate near discontinuous zones such as faults and unconformities. Moreover, the accuracy of an extracted horizon increases with increasing user interpreted control points. I improve the automatic seismic horizon interpretation from three aspects: (1) improving the accuracy of the seismic reflector dip attribute, (2) tracking a horizon using multiple seismic attributes, and (3) automatically generating control points prior the automatic tracking horizons. The extracted seismic horizons strictly follow the local seismic reflection events over the whole seismic survey. Automatic or semi-automatic fault surface construction is still a challenges task although seismic fault attributes are widely used in assisting seismic fault interpretation in 3D seismic survey. The staircase and undesired sequence stratigraphic artifacts are the main factors that hinder researchers from automatically constructing fault surfaces. I improve the automatic seismic fault interpretation from two aspects: (1) generating a new seismic fault attribute without staircase and undesired sequence stratigraphic artifacts, and (2) automatically constructing fault surfaces by analyzing the topological features of the new seismic fault attribute on time and vertical slices. The proposed fault surface construction workflow successfully constructs fault surfaces and computes corresponding fault parameters such as fault dip and strike and even conjugate faults within the seismic survey.