The mortality of single-motorcycle accident is significantly higher than other types of crashes involving motorcycle. In order to prevent the occurrence of single-motorcycle deaths, accident scenes are identified from the 3 aspects such as drivers, roads and environment. Firstly, the factors influencing the severity of single-motorcycle crash accidents are researched. Collecting 4,067 single-motorcycle crash items with 13 independent variables involving drivers' characteristics, road conditions and environment, binary logistic model was constructed to account for the correlation between the 13 factors and severity of accidents. The results show that month, weather, road types, road alignment, gender of motorcyclist, age of motorcyclist, alcohol use and helmet usage are associated with the death or injury at 0.1 significance level. Secondly, using the association rules based on Apriori algorithm, 10 different death scenes of single-motorcycle which includes characteristics of motorcyclists, roads and environment are identified. At the same time, a series of death prevention measures are proposed for the traffic management departments. The method of scene identification put forward in this paper can also be used as a theoretical basis for the prevention of similar traffic accidents.

CICTP