The present invention provides a method for determining the position of a moving object, comprising: Step a, obtaining a surround-view image of the surroundings of a moving object for each unit of time; Step b, determining, on the basis of the surround-view image, a virtual target area using a pre-trained recognition model; Step c, determining imaging position information of each virtual key point of the virtual target area for the virtual target area using the recognition model; Step d, computing, on the basis of a mapping relation between an imaging coordinate system of the surround-view image and a subject coordinate system of the moving object and according to the imaging position information of each virtual key point, a relative distance between an actual target area corresponding to the virtual target area and the moving object; and Step e, for each unit of time, correcting the relative distance obtained by means of Steps a to d at a current unit of time by using a difference between the relative distance obtained by means of Steps a to d at a previous unit of time before the current unit of time and a movement distance of the moving object in the previous unit of time, so as to obtain a corrected relative distance at the current unit of time.