Abstract: 【Objective】To explore the accuracy of classification of different models and the urbanization process of Nanjing during 2001-2015, we used Landsat 5 TM imagery acquired in 2006 and 2010, Landsat 7 ETM+ imagery acquired in 2001, and Landsat 8 OLI imagery acquired in 2015, covering Nanjing City. 【Method】We computed the modified normalized difference water index(MNDWI)to mask out water bodies within the study area, followed by the extraction of the built-up areas and non-built-up classes via five urban index models(UI, the urban index; NDBI, the normalized difference built-up index; IBI, the index-based built-up index; VrNIR-BI, the visible red-based built-up index; and VgNIR-BI, the visible green-based built-up index)based on a maximum inter-class difference algorithm derived from the Landsat images. 【Results】The validation results indicated that the accuracy of all models was over 64.33%. Among all six models, UI and NDBI had higher overall mapping accuracy, and their overall accuracy was among the highest(85.00% and 82.60%)in 2010, and the lowest(69.67% and 72.00%)in 2001, respectively. By calculating the increment of built-up area from these two indices, the urbanization degree calculated from UI increased from 11.55% in 2001 to 20.50% in 2015, whereas the urbanization degree derived from NDBI increased from 17.66% in 2001, to 25.60% in 2015. 【Conclusion】Furthermore, without considering Gaochun and Lishui Districts, the top four districts in urbanization processes of UI and NDBI were Jiangning, Luhe, Pukou, and Qixia.