Soil erosion by water is detrimental to soil fertility, crop yield, and the environment. For cold areas, knowledge of winter hydrologic processes is critical to determining land-use and management practices for reducing soil loss and protecting land and water resources. Adequate understanding of winter processes is also essential to developing models as effective predictive tools. This study evaluated the effects of two contrasting tillage practices on winter hydrologic and erosion processes, and the suitability of the Water Erosion Prediction Project (WEPP) model with a newly implemented energy-budget-based winter routine for quantifying these processes. Research plots subject to two tillage treatments—continuous tilled bare fallow (CTBF) and no-till (NT) seeding of winter wheat (Triticum aestivum L. cv. Madsen) after spring barley (Hordeum vulgare L.)—were established at the USDA-ARS Palouse Conservation Field Station, Pullman, WA. The plots were monitored for runoff, erosion, soil temperature, water content, and depths of snow and freeze–thaw during October to May of 2003–2004 through 2006–2007. The NT plot generated negligible runoff and erosion (0.5 mm, 0.2 Mg ha−1) compared with CTBF (323 mm, 547 Mg ha−1). Frost occurred more frequently and was deeper in CTBF, probably due to its lack of residue and shallower snow depth. The modified WEPP model could reasonably reproduce major winter processes, yet it cannot represent all the complicated winter phenomena observed in the field. Continued efforts are needed to further improve the ability of WEPP to properly account for soil freeze–thaw and thus transient soil hydraulic properties and hydrologic and erosion processes.