Among soil pollutants irresolvable heavy metals, due to their toxicity, accumulation properties and physiological effect on living in low concentration, are of most importance. These elements due to their low mobility accumulated in soils in time. Entering the food cycle, their accumulation will thread human beings as well as animals. The heavy metals accumulate mostly in the surface soil, and thus exposure more to plant uptake and traort. Thus studing concentration distribution for soil pollution monitoring of heavy metals is indispensible for maintaining the environment quality and should be considered at different levels of dicision making by researchers and managers. The aim of study is the evaluation of spatial distribution of Cu and Zn in the surface soil of Hamadan Province with classification and regression tree (CART) model and compare with kriging, assessment of soil pollution of these metals using geoaccumulation index, pollution index and integrated pollution index. For this purpose, the surface soil samples (0-20 cm) were taken using systematic random sampling method and the concentration of heavy metals and some soil parameters such as pH, EC, the percentage of organic matter, calcium carbonate, calcium sulfate, sand, silt and clay were analyzed. Spatial distribution maps of these parameters were made by comparison of accuracy of different kriging methods, Radial Basis Function (RBF) and Inverse Distance Weight (IDW) by the Mean Absolute Error (MAE) values, Mean Bias Error (MBE) and Root Mean Square Error (RMSE) and then primary dataes were extracted for preparing CART model. In addition, needed dataes were extracted by elevation of sea level, slop, geographical aspect, geology and land use maps of Hamada Province and final CART model was prepared. Nevertheless, spatial distribution map of Cu was prepared by ordinary kriging method and exponential model and spatial distribution map of Zn was prepared by simple kriging method and gaussian model for assessment of kriging method accuracy compared with CART model. The regions of western and northern Hamadan Province respectively had the highest and lowest concentration of Cu and Zn. The most important factors to concentration of Cu following results of prepared model can indicate slope, calcium sulfate and calcium carbonate, elevation and percentage of organic matter and most important factors to concentration of Zn, percentage of calcium carbonate, slop, sand, silt and land use were determined. The results indicate that accuracy of CART model is higher than kriging method using coefficient of determination (R 2 ) and Mean Square Error (MSE) indices. The assessment of soil contamination using geoaccumulation index (IGEO) based on the background of concentration in the study area, classified the contamination as low to negligible in the region. According to the pollution index (PI), more than half of the samples had intermediated pollution based on Cu and Zn. According to the integrated pollution index more than 70% of the samples, classified the contamination as medium pollution based on these metals. Due to consistency of he areas with high concentrations of the elements with agriculture landuse, the Integrated Pollution Index (IPI) indicated that agriculture activities could also increase the concentration of these elements in the soils