The height in 500 hPa was used as the key influencing factor,using K-means dynamic clustering method,the snowfall situations in Anshun was divided into three main situations: straight airflow pattern (Ⅰ), south branch trough pattern (Ⅱ), and multi-wave pattern (Ⅲ). Case analysis and synthetic analysis show that the characteristics of the phase change, region, and duration of the snow fall in these three situations were different. Synthetic analysis shows that the three situations have obvious differences in the 500 hPa height： pattern Ⅰ cases were basically straight from the east of the plateau to Guizhou, pattern Ⅱ had an obvious south branch trough from the east of the plateau to Guizhou, and the pattern Ⅲ had multiple wave troughs from the east of the plateau to Guizhou, forming a step troughs. In addition, the characteristics of Yunnan-Guizhou quasi-stationary front strength influence area, physical quantity, vertical humidity configuration and temperature inversion, etc. were different:the wet layer and the uplift zone were shallow in pattern Ⅰ , with temperature inversion around 1℃; the wet layer was thinner in pattern Ⅱ,having a narrow range of 3℃ temperature inversion in the middle of Guizhou;in the middle and low layers there was a wide wet layer, and water vapor and uplift conditions were well coordinated in pattern Ⅲ , there was almost no inversion temperature.