The study aimed to model earth land use changing using logistic regression and multi-layer Perceptron artificial neural network in the South Coast of the Caspian Sea in Iran. Satellite Landsat images belonging to 1993, 2000, 2008 and 2017 has used. Modeling the potential of earth covering change was conducted using logistic regression and multi-layer Perceptron artificial neural network. Markov chain and a hard prediction model was used to forecast earth covering changes from 1993-2017 calibration method. Kappa coefficient results showed that regression (0.8657) has more accuracy compared to a multi-layer Percept.
