A vehicle periphery monitoring device which determines type of an object with high accuracy, wherein the size of object regions that are set in each of the image representing the peripheral condition of the vehicle in each of two different points in time and that include the identical objects, are aligned on the basis of the distance from the vehicle to the object in each of the two different point in time. Further, local regions with the same arrangement pattern are set taking each of the object regions with aligned size as reference. Still further, the object is classified into the object class which corresponds to the arrangement pattern in the case where the degree of correlation between the local regions becomes equal to or larger than a threshold value.