Soil water repellency induced by wildfires can alter hydraulic properties and hydrologic processes; however, the persistence and vertical position (i.e., depth) of water repellent layers can vary between systems and fires, with limited understanding of how those variations affect infiltration processes. This study occurred in two forested locations in the south‐central Appalachian Mountains that experienced wildfires in late 2016: Mount Pleasant Wildfire Refuge, Virginia, and Chimney Rock State Park, North Carolina. In each location, sites were selected to represent unburned conditions and low to moderate burn intensities. At each site, we measured the soil water repellency at the surface (ash layer or O horizon) and ~2 cm below the surface (A horizon) using the water drop penetration time (WDPT) method (n = 10-14). Soil water content was also measured over the upper 10 cm of the soil (n = 10), and infiltration tests were conducted using a tension infiltrometer (n = 6‐8). The results showed that soil repellency was highest in the surface layer at the Mount Pleasant location and was highest in the subsurface layer at the Chimney Rock location. Soil water content was lower in unburned soil than in burned soil, especially for measurements taken immediately post‐fire, with soil water content negatively correlated with water repellency. Water repellency in the surface layer significantly reduced relative infiltration rates (estimated as differences between initial and steady‐state rates), whereas subsurface water repellency did not affect relative infiltration. As a result, water repellency persisted longer in sites with surface as opposed to subsurface water repellency. Finally, differences between burned and unburned sites showed that while the wildfires increased the occurrence of water repellency, they did not alter the underlying relationship between relative infiltration and water repellency of the surface soil.