Abstract: Based on highway traffic accident data from 2010 to 2015 in Liaoning province,including geographic position,level of accident and environmental condition,highway traffic accident index was quantified based on the level and the number of accidents using order relation analysis method.By matching the accident location with the nearest meteorological observation station,correlation relationship between meteorological factors and traffic accident number was analyzed.A stepwise regression analysis was proposed to estimate highway traffic accident index based on meteorological factors.The results show that the years with high frequency of highway traffic accidents are characterized by more precipitation and less sunshine hours.The number of accidents caused by weather and average highway traffic accident index are the largest in winter.The correlation relationship between highway traffic accident index and relative humidity (or precipitation) is better than that with other meteorological factors.The stepwise regression method is used for simulating highway accident index based on meteorological factors.It suggests that the prediction accurate rates of major accidents in winter and autumn are higher than those in other seasons,and they are 62.5% and 51.3%,respectively.