Soil moisture is a critical variable in the water and energy cycles. The prediction of soil moisture patterns, especially at high spatial resolution, is challenging. This study tests the ability of a land surface hydrologic model (Flux‐PIHM) to simulate high‐resolution soil moisture patterns in the Shale Hills watershed (0.08 km2) in central Pennsylvania. Locally measured variables including a soil map, soil parameters, a tree map, and lidar topographic data, all have been synthesized into Flux‐PIHM to provide model inputs. The predicted 10‐cm soil moisture patterns for 15 individual days encompassing seven months in 2009 are compared with the observations from 61 soil moisture monitoring sites. Calibrated using only watershed‐scale and a few point‐based measurements, and driven by spatially uniform meteorological forcing, Flux‐PIHM is able to simulate the observed macro spatial pattern of soil moisture at ∼10‐m resolution (spatial correlation coefficient ∼0.6) and the day‐to‐day variation of this soil moisture pattern, although it underestimates the amplitude of the spatial variability and the mean soil moisture. Results show that the spatial distribution of soil hydraulic parameters has the dominant effect on the soil moisture spatial pattern. The surface topography and depth to bedrock also affect the soil moisture patterns in this watershed. Using the National Land Cover Database (NLCD) in place of a local tree survey map makes a negligible difference. Field measured soil type maps and soil type‐specific hydraulic parameters significantly improve the predicted soil moisture pattern as compared to the most detailed national soils database (Soil Survey Geographic Database, or SSURGO, 30‐m resolution).