Haftad-Gholleh Area has recently encountered many changes. In this study, by using landsat images of the years 1996 and 2016, landuse maps were classified into four classes including: agriculture, rangeland, residential areas, and rocks. Land Change Modeler (LCM) and Habitat and Biodiversity Modeler (HBM) modules in the Idrisi GIS software were used to analyze the land use changes and habitat evaluation for the prediction of the land uses status in 2016, based on the Artificial Neural Network (ANN), Markov Chain analysis and logistic regression. The results showed that most of the changes in the landscape of the region between 1996 and 2016 were related, respectively, to attrition, aggregation and creation indicators; between the years 2016 to 2041, they can be related, respectively, to the creation and dissection indicators. The habitat evaluation showed that 4.5% of the habitat was decreased in 2016, as compared to 1996. With the continuation of this trend, 6.5% of the habitat will fall in 2041, as compared to 2016. Receiver Operating Characteridtic (ROC) of the model also specified that the desirability model validity was equal to 0.9558, showing the excellent performance of logistic regression method. In general, this can be an important principle approach preventing from changes in the habitat of wild sheep to other land uses.