Abstract: To study the snowfall characteristics in Qingdao and the formation causes, the NCEP global reanalyzed grid data and observational data of 1961-2009 were employed to analyze the temporal and spatial characteristics of snowfall in Qingdao. Using the method of composite analysis and physical diagnosis, the circulation patterns and physical characteristics of the Qingdao heavy snowfall were obtained. The results showed that: the earliest snowfall in Qingdao was in November, the latest was in April; the average annual snowfall was around 10 mm, more in the northeast and southwest, and less in the southeast in Qingdao; the average annual snowfall days were about 10 days, more in the northwest, and less in the middle; the monthly average snowfall and snowfall days showed the monthly variation characteristics of high in the middle and low at both ends, more in December, January and February, and less in November and April; the wavelet analysis results showed that both annual snowfall and annual snowfall days had the characteristics of oscillations with a quasi-19 year period; heavy snow first appeared in late November and the latest was in the middle of March. There were three kinds of circulation patterns for heavy snow in Qingdao, which were cold-trough eastern moving pattern, straight circulation pattern and zonal trough shifting to meridional pattern, and the main pattern was cold-trough eastern moving pattern; there were obviously cold air in the upper air and confluence of cold and warm air on the ground in Qingdao, meanwhile the warm air carried abundant moisture and the humidity increased. Qingdao was in the divergence area in high-level and convergence area in low-level. Such circumstances caused significant snowfall weather in Qingdao. Thus, physical characteristics analysis could enhance the heavy snow forecasting in Qingdao.