Rapid land use change has taken place over the last few decades in Istanbul. As most of the metropolitan areas, Istanbul faces increasing problems connected to increasing population and urbanisation. In this study, temporal changes of Istanbul's land use/cover were defined using remotely sensed data and post classification change detection method. For the aim of the study, relevant information was derived from different dated Landsat Thematic Mapper (TM) satellite data by using Unsupervised Iterative Self-Organizing Data Analysis Technique (ISODATA) and results were examined with matrix analysis method. Ground truth data were used for the classification and accuracy assessment of the classification. Temporal changes of land use/cover classes of the mega city Istanbul between the years of 1992, 1997 and 2005 were examined for the management and decision making process. Landsat TM images were classified into six land use/cover types: forest-green area, bare land, water surface, road, urban area, and mining area. The results show that urban areas and road categories are increased greatly by 13,630 and 5,018ha, respectively, but forest-green areas decreased by 77,722ha over 13years between 1992 and 2005. The reason for the decrease in green areas is mainly because of development of unplanned urbanization and unavoidable migration.