The lack of quantitative information on plant species distribution and their response to environmental factors, lack of species distribution map on long to, advances in the environment layers production (GIS, remote sensing and statistics), caused species distribution modeling become more important in recent decades. The current research aims to modeling and mapping potential habitat of Astragalus verus using the Classification and Regression Tree and Nonparametric Multiplicative Regresson in western Isfahan province in an area by 6,123,656 hectares. In order to sampling on occurrence of Astragalus verus species, a stratified random sampling method in 287 sites (including 106 presence site and 181 absence site), was used to record the occurrence of species. Environmental maps were developed using geostatistical methods (kriging and inverse distance weighting) using Arc GIS software for the whole study area. The relationship between the occurrence of given species and different environmental factors were determined using two statistical methods CART and NPMR. produced models were converted from ecological space into geographical space using GIS. Results of NPMR model showed that the presence of species has the highest correlation with environmental variables, such as the mean annual temperature, seasonal temperatures, the maximum temperature of the warmest month, and the average temperature in the coldest season, rainfall in the driest months, rainfall in the coldest season and soil acidity. Additionally, results of PCA analysis to select the input variables for CART model showed that eight environmental variables (the average temperature in the dry season, the minimum temperature in the coldest month, average temperature in the coldest season, mean annual temperature, precipitation in the driest season, organic carbon, soil content clay and saturation moisture) accounts for the most important environmental factors in the spatial distribution of species Astragalus verus. In another approach to compare the two models CART and NPMR, input variables for both models were considered the same. The results of the model indicate that performance of both models CART and NPMR with area under curve ROC about 0.94 and kappa coefficient 0.91 and 0.72 and 0.74 are excellent and good respectively. These results indicate high potentials of models to generate species distribution maps in Astragalus verus. Using the optimum threshold from models validation, habitat suitability map (zero or one) for species Astragalus verus was obtained. Suitable habitat for the species level according to the models CART and NPMR equaled to 2115760 and 1,849,269 hectares (34.59% and 30.72% of the total area) respectively. by reviewing the CART model and output response curves generated by the model NPMR it was found that Astragalus verus showed the highest presence in upland habitat (from 2900 to 1800 m), areas with slopes less than 12.5% the average annual precipitation 125 mm and a mean annual temperature between 10-14 ° C and soils with pH neutral to slightly alkaline (7.8 = PH) and soils texture moderate to heavy (between 22 and 44%clay) and the relatively low permeability. The results showed that both the CART and NPMR models may be promising in the planning and management of sustainable development, restoration, conservation, evaluation and rehabilitation of rangeland as well as mapping the distribution of species and species response curves to environment variables.