Metropolitan cities in developing countries face dynamic urbanization due to population expansion and migration. Continuous construction of urban components related to urbanization progress result with important changes in land cover and land use characteristics of the city. Such changes generally have negative effects on the natural habitat and ecosystem. Thus, urbanization related changes should be monitored and current situation should be determined in regular intervals. Istanbul is a metropolitan city that exposed to continuous urbanization with the prominent position that connects the European and Asian continents economically and culturally. Although, it is the most crowded city of Europe with a population of around 17 million, Istanbul faces a continuous and fast urbanization progress. Main scope of this study is to determine the changes occurring in land cover due to the rapid increase of urbanization within the provincial borders of Istanbul and to present the current situation. In this context, object based classification has been applied to built up index dataset produced from multitemporal Landsat 8 satellite images that were acquired between 2013 – 2017. Within the scope of the analysis, the built up index data were produced for each year after the preliminary processing steps including data fusion and radiometric calibration processes for satellite images were completed. In the next step, segmentation was applied to index dataset in order to produce the objects. As a further step, decision tree based classification was applied to dataset according to class labels that were defined as thresholds derived from the spectral and textural properties of objects. As a last step, accuracy assessment was applied to classification result. Results of the study indicates that current land cover situation and temporal changes can be determined with accuracy around 90% by the proposed method.