Abstract: The static data of WRF model is refined by latest topographic data and land cover data. The ERA-Interim data is used as the initial field and boundary field, and the hourly visibility observation of 84 national stations in Guizhou area is used as actual situation, the effect of the refined WRF static data on the simulation of quasi-stationary front frontal fog is analyzed through comparison experiments. It is indicated that the improved topographic data reduces the absolute altitude deviation of 52.37% stations in Guizhou, and the new land cover data can describe the spatial distribution of land-cover in Guizhou more accurately. The individually improved land cover data has little effect on the simulation. The individually improved topographic data improves most of the simulated results, and the topographic data improvement effect is proportional to the simulation improvement effect. Refining the topographic data and land cover data at the same time, the number of improvement cases and the average improvement rate decreases slightly, but the average false alarm ratio performs significantly better than that of only refining the terrain. For different types of quasi-stationary fronts, the WRF model not only has different simulation capabilities, but also has different improvement effects on static data. The refined static data has a mostly positive effect on the simulation results to quasi-stationary fronts when the front area maintains strong strength and wet area changes little. When the front area oscillates east-west, and the wet area has great changes, especially in the wet area with RH > 95% in the surface layer, refined static data only has little positive effect on the simulation results to the quasi-stationary fronts. Based on the current distribution of visibility observation stations, improved topographic data and land cover data can positively improve the quasi-stationary front frontal fog simulation in Guizhou.