Water management in Chennai holds significance for urban sustainability, with sources such as Chembarambakkam Lake and Poondi Reservoir playing a major role in supplying water for the city. This indicates that drainage connectivity and its analysis are inadequate and lead to substantial flooding; thus, more segmentation and management of drainage pathways are needed to control water flow and intensity of flooding. This study examines Chennai as a case study using a 30m-resolution Digital Elevation Model derived from Shuttle Radar Topography Mission data to capture the region’s elevations. This research processes the Digital Elevation Model data through flow direction and accumulation analysis, generating a Topographic Wetness Index image. The Topographic Wetness Index image assists in improving the elevation data by reducing the elevation of the water bodies while increasing the land height surrounding the water bodies, which assists in defining the drainage network boundaries. To segment the drainage patterns in detail, the article utilizes a Vision Transformer that uses enhanced Topographic Wetness Index images for training and testing. The architecture of the Vision Transformer is such that the recognition of complex patterns is made possible, thereby making way for accurate segmentation and mapping of drainage networks. The model’s performance was successfully delineating and analyzing the drainage networks identified in Chennai, with a 98.64% accuracy for training and 98.94% for testing using Digital Elevation Model data. This thus shows the potent ability of the Vision Transformer as a tool for the hydrological studies of urban environments.