The aim of the study was to examine the potential maximum likelihood classification in the mapping of basic land cover/land use classes by using ALOS AVNIR-2 imagery. The two specific objectives were; (a) to develop a maximum likelihood classification scheme for mapping land cover/land use classes using ALOS AVNIR-2 imagery, (b) to estimate the accuracy of the used method. Land cover nomenclature is classified according to the Coordination of Information on the Environment (CORINE) Level 2 and 3 classifications. Ten urban land cover classes were used in this study: river, wetland vegetation, forest, mining area, shadow, mountain forest, cemetery, agriculture, built up area, industrial area. The classification accuracy was assessed using 218 pixels were stratified randomly distributed throughout the study area and independent of training sites used by the supervised classification algorithm. The results show that overall classification accuracies is 81.19% and overall kappa statistics is 0.7845.