Summary:　In　this　paper,　we　present　a　structure-adaptive　anisotropic　Gaussian　filter　to　denoise　and　enhance　3D　reflection　seismic　data　without　blurring　relevant　details　such　as　structural　and　stratigraphic　discontinuities.　This　filtering　framework　relies　on　the　computation　of　the　Gradient　Structure　Tensors　(GST)　providing　information　on　the　local　orientation　of　the　geological　structures.　The　filter　kernel　is　anisotropically　constructed　based　on　two　confidence　measures,　both　of　which　take　into　account　the　regularity　of　the　local　seismic　structures.　So　that,　the　filter　shape　is　well　adjusted　in　accordance　with　different　local　geological　features.　Then　the　filtering　process　is　steered　according　to　the　directional　information　provided　by　the　GST.　We　demonstrate　the　efficiency　of　the　proposed　filter　through　both　a　comparative　experiment　with　the　well-known　＂Structure-Oriented　Filter-Edge　Preserving＂　(SOF-EP)　on　synthetic　blocks　and　an　application　to　real　3D　seismic　data.