The aim of this study was to assess the application of cellular automata in urban modeling to give insights into a wide variety of urban phenomena, using the most commonly used urban modeling approaches including: Markov Chain, SLEUTH, Dinamica EGO modelling with the Logistic Regression (LR), Regression Tree (RT) and Artificial Neural Networks (ANN). The effectiveness of these approaches in forecasting the urban growth was assessed in the example of Adana as a fast growing City in Turkey for the year 2023. Different models have their own merits and advantages, the empirical results and findings of various approaches provided a guide for urban sprawl modeling. The accuracy figures to assess the models were derived using Allocation and Disagreement maps together with Kappa statistics. Calibration data were from remotely sensed images recorded in 1967, 1977, 1987, 1998 and 2007. SLEUTH, Markov Chain and RT models resulted in overall Kappa accuracy measures of 75%, 72% and 71% respectively, measured over the past data using hindcasting. LR and ANN yielded the least accurate results with an overall Kappa accuracy of 66%. Different modeling approaches have their own merits. However, the SLEUTH model was the most accurate for handling the variability in the present urban development. (C) 2016 Elsevier B.V. All rights reserved.