The confirmation of the driving speed of vehicles in traffic accidents is an important basis for analyzing the causes of acci-dents and determining the responsibility.In the process of handling traffic violation cases,the amount by which the vehicle's speed ex-ceeds the prescribed limit on the road also determines the severity of the punishment imposed on the driver.When calculating the driv-ing speed of a vehicle using fixed video images,it is usually necessary to select two feature points that are equidistant from the longitu-dinal symmetry plane of the vehicle at a height equal to the ground on the same side of the body surface of the target vehicle.The dis-tance between the two points is the driving displacement of the identified vehicle.When the video image is blurry and unclear,and it is difficult to select the aforementioned feature points according to standards,only relying on the easily recognizable feature positions of the vehicle body on the same side can make a rough estimation of the vehicle's driving speed,and the accuracy of the calculation results is relatively low.By establishing a mathematical model,analyze and study the driving speed of a vehicle in a traffic accident,adjust rel-evant parameters to improve the accuracy of the identification results.