The purpose of this study was to survey and model the land use change in Khorramabad city. For this purpose, using three Landsat satellite images of years 1986, 2001 and 2016, the cover and land cover classes have been extracted from Khorramabad city and around it. Overall classification accuracy for image corresponding to the year 1365, 93/89, 1380, and 34/91 of 1395, 62/95 percent is obtained. In order to model the land use changes in Khorramabad, the input layers of elevation, slope, shadow, distance from the road, distance from the built area, distance from the agricultural land, distance from the forestland and the distance from the mountainous terrain were used. In the following, neural network models and CA-MARKOV are used to model and predict land use changes by the year 1404. The results show a high accuracy of 65% for modeling land use change in Khorramabad city. It should be noted that this precision is reasonable for predicting and modeling land-use changes that is dynamic, because in addition to the variables considered in this research and other studies, other factors such as municipal regulations, The land stock exchange, state-owned large-scale housing policies, and so on, will have an impact on land use change.