Abstract: Using conventional observation data, ground automatic weather stations data, NCEP 1°X1° reanalysis data, cloud top temperature (TBB) data provided by the National Satellite Meteorological Centre, and dual-polarization Doppler weather radar products, the extreme snowstorm process in Hunan from February 22 to 23, 2022 was analyzed. The results show that: (1) A medium-β-scale cloud cluster embedded in the southwest-northeast cloud belt that maintains in the mature stage continues to pass through to form a propagation similar to the train effect, which was the main reason for this extreme snowstorm process, the extreme snowstorm area was mainly located at the edge of the cold cloud cluster with TBB<30 ℃. (2) The correlation coefficient, reflectivity factor and differential reflectivity of the dual-polarization Doppler weather radar had good indication for the identification and judgment of the height of the melting layer and the properties of precipitation particles in the atmosphere. The correlation coefficient was 0.8~0.97, which basically corresponded to the characteristics of the melting level. (3) The VWP product had a certain indication for the phase transition of rain and snow and the strengthening of snowfall. The VWP products could detect a stable vertical wind shear formed at 2.4 km height before snowfall, and the enhancement of the northeast jet at 1.2 km height, the enhancement of the southwest jet at 3 km height, and the increase of the vertical wind shear at 2.4 km height during the snowfall enhancement. The stable vertical wind shear at 2.4 km height stabilized for more than ten hours, which provided strong dynamic conditions for the extreme snowstorm process and was conducive to the establishment and maintenance of unstable stratification. The rapid conversion of rain and snow and the large snow depth in this process were mainly caused by the relatively dry and cold lower layer in the early stage of snowfall.