The invention provides a traffic event detection and positioning method. According to the invention, statistics of lane switching feature parameters is performed according to a specific time interval by taking lanes as the unit and then whether traffic events occur in a lane or not is judged by adopting a threshold value method on the basis of calculating prior vehicle switching-out rate of the lane. As for the lane judged to have traffic event, a switching-in point and a switching-out point of lane switching of the lane are taken as a positive sample and a negative sample. Accurate computation of event point positions is realized through calculating the optical classification plane of the two groups of samples. According to the invention, an event point positioning problem is creatively turned to a sample classification problem of a lane switching point and an accurate positioning problem of the traffic events is solved from the perspective of lane switching features. The method has advantages of short detection time, high positioning precision, high information transmission speed and the like. The method can provide powerful support for timely detection and treatment of traffic events and prevention of secondary accidents.