The main objectives of this resaerch were to study the topographical effects on soil characteristics and wheat production, and compare the statistics, intelligent and dynamic models to prediction wheat production in the central Zagros of Iran. The results showed that, topography and slope position had significant effect on both soil characteristics and wheat yield. Soil depth modeling showed that, slope was the most predictor. Wheat grain and biomass yield had significant correlation with saturation percentage and soil organic matter. Slope position had significant effect on grain and biomass wheat yield. The best suitable position was toeslope and the worst was shoulder for wheat production. In second study, among the MLR, SVM and ANN models, the ANN models were the best to predict wheat grain and biomass yield. Sensitivity analysis by Hill method showed that, the most important factors affecting grain and biomass of rainfed wheat, were those that controlled soil water content. The results of prediction of wheat yield by CERES-wheat model showed that, the model doesn’t have any perception about the hillslope or any of the slope positio actually the model’s performance is satisfactory in those slope positions which they have the most analogy with the ideal assumed field by DSSAT package. Keywords : modeling, slope position, wheat yield, intelligent models, DSSAT, Zagros
