Soil erosion is one of the most important economic, social and environmental problem in the world. Soil erosion is the main index in ecosystem health assessment. According to available information, the Asia suffer more than other continents, in soil erosion. In our country due to lack of sufficient number of measuring stations and lack of long-term data bases, for estimating soil erosion in watershed, hydrological modeling is a main concern for evaluating soil erosion in watershed scale. Modeling as a powerful procedure has been used sucessfuly for more tha half century for predicting soil erosion. In the present study the SWAT2005model and land capability evaluation were used to land use planning in Zayanderoud watershed to achieve reduction of soil loss.The SWAT model is a continuous, conceptual and long-term field scale model that was being designed based on daily and sub-daily periods. Calibration, validation and uncertainty analysi have done using SUFI-2 program. After collecting input data including digital elevation model (DEM),strearm network, land use, soil map, database files, climate and soil temperature and rainfall data bases for minimum and maximum meteorological station study area model was ran for period of 1987-2005. 75 percent of data were used for calibration and 25 percent for validation. In order to evaluate efficency of model, the validation indices such as p-factor, d-factor, coefficient of determination (R 2 ) and t Nash - Satclife coefficien(NS) were applied. Results of monthly runoff simulation was very satisfactory, so that R 2 and N-S were 0.79 and 0.69 for calibration and .075 and 0.61 for validation respectively. For daily runoff and daily sediment simulation R 2 and N-S calculated 0.75 and 0.67 for calibration daily runoff, and 0. 57 and 0.51for calibration daily sediment respectively, and 0.72 and 0.59 for validation daily runoff, 0.52 and 0.45 for validation daily sediments respectively. P-factor in all calculation steps caculated larger than 50 % and d-factor calculated less than 1 was also conifremed the satisfaction in runoff and sediment yield simulation. Then the land capability evaluation of the selected watershed was persormed using Mahler approach and the best relevant of major land use was determined for each of land units. Afterwards, the calibrated and validated SWAT model was ran for the proposed land uses as a new scenario. Finaly the resultes obtained from the two scenarios (current and proposed land uses) were compared using t-test computation in 99% probability level. The resultes of comparision of two mentioned scenario revealed Keyword s: SWAT2005 model, land evaluation, calibration, validation and Nash-satclife coefficient.
