A　novel　bond-based　nonlocal　fluid　transport　model　is　introduced　to　tackle　fluid　flow　in　saturated　porous　media　characterized　by　anisotropic　and　heterogeneous　permeability.　This　model　is　built　upon　the　bond-based　framework　and　utilizes　a　type　of　Gaussian　kernel　function　with　truncation,　incorporating　the　coefficient　matrix　within　the　covariance　matrix.　It　is　crucial　to　highlight　that　the　permeability　exerts　a　profound　influence　on　the　size　and　shape　of　the　influence　region　after　truncation.　As　a　result,　in　cases　of　strong　heterogeneity,　the　calculation　area　can　become　excessively　large,　posing　challenges　to　effective　computation.　To　confront　this　difficulty,　we　introduce　a　reconstructed　bond-based　nonlocal　fluid　transport　model　that　dynamically　adjusts　based　on　permeability,　allowing　it　to　accommodate　a　wider　range　of　scenarios.　Meanwhile,　we　have　proved　that　the　proposed　operator　convergence　to　the　local　fluid　transport　operator　in　weak　form.　Moreover,　a　fully　implicit　algorithm,　integrated　with　the　Newton-Raphson　method,　is　utilized　to　tackle　the　complex　nonlinear　issues　arising　in　saturated　porous　media　characterized　by　anisotropic　and　heterogeneous　permeability.　The　novel　bond-based　model　surpasses　traditional　bond-based　model　by　enhancing　fluid　transport　simulation　in　saturated　porous　media　with　strongly　heterogeneous　permeability.　Additionally,　unlike　non-ordinary　state-based　model,　the　proposed　model　eliminates　the　need　for　complex　penalty　methods　to　address　zero-energy　modes,　significantly　reducing　computational　demands　and　time,　as　demonstrated　by　further　validation.　The　numerical　example　exhibits　strong　agreement　with　COMSOL　Multiphysics,　confirming　the　accuracy　of　the　model　and　its　capability　to　effectively　capture　pressure　discontinuities　along　material　interfaces　and　crack　surfaces　in　saturated　porous　media　characterized　by　anisotropic　and　strong　heterogeneous　permeability.　©　2025　Author(s).