Most land surface models (LSMs) do not explicitly represent surface ponding, infiltration of ponded water, or the soil macropore effects on infiltration, percolation, and groundwater recharge. In this study, we implement a dual-permeability model (DPM) based on the mixed-form Richards’ equation, which solves pressure head continuously across unsaturated and saturated zones while conserves mass, into the Noah-MP LSM to represent slow flow through soil matrix and rapid flow through macropore networks. The model explicitly computes surface ponding depth, infiltration of ponded water, and runoff beyond a ponding threshold (infiltration-excess runoff) by switching the atmospheric boundary condition between head and flux boundary conditions. The new model also provides two optional soil water retention models of Van Genuchten (VG) and Brooks-Corey (BC). Model experiments over the conterminous US indicate that 1) surface ponded water and its runoff contribute substantially to seasonal variations in total water storage and peak flows in wet regions with low soil permeability (e.g., the Lower Mississippi River and surrounding regions), 2) the VG model produces drier topsoil with less soil surface evaporation than does the BC model with the Clapp-Hornberger parameters, especially during droughts and in dry regions, better matching remote sensing soil moisture, and 3) DMP produces more runoff with increased subsurface runoff, thereby improving the modeling skill at monthly scale over all subbasins of the Mississippi River, especially for low flow events. This study also highlights the importance of consistent representations of soil and plant hydraulics in Earth System Models to modeling ecosystem drought resilience.