In this method, the land cover classes were classified one after another and in a predetermined order. The main strategy in the design of this arrangement is based on higher separability and class significance. Given the spectral characteristics of vegetation classes in the visible bands, vegetation classes have the best conditions for classification and so were classified first. After classification of the vegetation class, its objects were reclassified into tree and grass classes. By merging tree neighbor objects, this class was reclassified into single tree and garden tree (cluster) classes. The building objects, using the features based on the DSM, were classified in the next stage. The buildings, based on the DSM features, were reclassified into one, two, or more than two story (level) classes. After these stages, the road classes (considering the class importance criterion) were detected. Road class objects were reclassified into main and secondary road classes. Considering the spatial and conceptual correlation between car and bus elements with roads, the objects of these two classes were classified in the next stage. The shadows have higher separability between the remaining classes. The bare soil and open space class with high spectral heterogeneity and any geometrical attributes were classified at the end. According to the significance measure, the road class was extracted after these two classes. Vehicle and shadow classes were classified in the next stage. The bare soil and open space class with its spectral diversities, which have the least geometrical and conceptual attribute, were classified in the final stage in this model.