This study addresses the prevention and control requirements for windblown snow hazards on high-altitude roads in the southeastern Tibetan region. Based on the morphological characteristics of typical road cross-sections, an integrated approach combining field surveys and computational fluid dynamics simulations was employed to systematically investigate the mechanisms by which terrain slope and embankment excavation angle affect snow accumulation distribution. Within the framework of an Eulerian–Eulerian multiphase flow model, a numerical model for wind–snow two-phase flow was established. By comprehensively analyzing parameters such as wind velocity field, wall friction velocity, and snow contour lines, the distribution patterns of snow accumulation under various terrain conditions were examined in depth. The results indicate that variations in terrain slope significantly alter the structure of the flow field. When the slope exceeds 15°, flow separation is prone to occur at the windward crest, resulting in the formation of a leeward vortex region where snow particles predominantly accumulate; as the slope gradually decreases, the vortex region migrates toward the slope foot and eventually disappears at a critical slope of 15°. Furthermore, the extent of snow accumulation on the slope exhibits a nonlinear decreasing trend, with slopes of 45°, 30°, and 21.8° corresponding to snow coverage of approximately 7/8, 3/4, and 1/2 of the slope length, respectively. In addition, artificial excavation of the road embankment induces a geometric discontinuity that distorts the flow field at the slope foot, thereby causing snow to migrate toward the windward slope foot and significantly elevating the snow height. The primary affected area is approximately 10?m behind the excavation zone, and within this region the amount of accumulated snow decreases with increasing excavation angle. These findings provide a quantitative theoretical basis for the wind–snow resistant design of road engineering in high-altitude cold mountainous regions.