Slums are considered urban threats due to their overcrowding population, inadequate housing, and services. The social and physical growth of these slum areas is ignored due to the inadequate laws and management policies of urban bodies. From this end, the reason behind the underdevelopment is the absence of sufficient slum maps for the assessment of slums and proper management. In this regard, plenty of methods exist to solve this problem using remote sensing data. In modern days, deep machine learning (DL) approaches in remote sensing are being observed as an important tool for understanding the fundamental structure of physical characteristics situated in the images taken from satellites. This chapter tries to identify the unrecognized slums by using the random forest approach and also to predict the slums by applying cellular automata-Markov chain (CA-MC) techniques. For this, this study has collected training samples using Google Earth, and it is further used to delineate the slum boundaries in Hyderabad. The final output is validated concerning the original slum boundaries of Hyderabad city. The results of the validation process point to a classification process with more than 84.70 percent accuracy. This study has presented a detailed methodology for the delineation of the informal slums in the cities.