Aimed at the problem of inaccurate statistics caused by interference and noise data during a vehicle mileage statistics process, a vehicle mileage statistics method based on adaptive distance algorithm was given in this paper. By using the continuous satellite positioning data of specific vehicle, all satellite positioning data could be divided into low speed mode and high speed mode according to a predetermined speed threshold. An adaptive distance algorithm is developed, so the drift noise can be easily separated by the algorithm in real time. With some validations of real vehicle running data, compared with the existing algorithms, the algorithm developed in this paper has a good anti-interference ability to the satellite positioning noise data generated by the vehicle running conditions of parking and low speed. The influence of the drift noise on the vehicle mileage statistics can be effectively eliminated. At the same time, the algorithm can automatically filter the satellite positioning data whose latitude and longitude beyond the reasonable coordinate range. Experimental data shows that the algorithm can effectively improve the accuracy of vehicle mileage calculation.