Population growth leads to increase the food demands and put more pressures on natural resources especially the rangeland ecosystems. Rangelands are widely modified into rain fed agricultural fields to enhance the food production in the world. Inappropriate land use causes severe degradation and loss of vital resources such as water and soil nutrients in brittle rangeland ecosystems. Once the lands are degraded, the restoration programs are costly and time consuming; therefore the management must focus to prevent these undesirable changes. Achieving a sustainable management in rangeland and agricultural ecosystems requires recognizing the land capabilities and conducting a comprehensive assessment procedure to prevent soil erosion and crossing the functional thresholds in these ecosystems. Several models and tools such as Musgrave, EPM, MPSIAC, FAO and etc were developed to assess and monitor the soil erosion around the world. These models are calibrated based on the specific conditions of the areaes, where the models were developed. Although, some of these models are widely used in Iran and had relatively acceptable performances, their results must be verified and validated by direct and field works to rely on the obtained results. This study aimed to evaluate the application of Land Use Impact Model (LUIM), a relatively new method for assessing and monitoring soil erosion risks, in Karchambooy watershed basin – Isfahan. The LUIM incorporate the relationships between landscape characterization and land management practices to map the soil erosion risk in the study area. Bayesian Belief Networks () was used to link these data into a risk assessment framework in LUIM. Follow up workshops were used to elicit expert knowledge about the effects of various management practices on the soil erosion risk. The collected biophysical data from the study area were then incorporated with expert knowledge using within the LUIM. According to the results, rangelands had high erosion probability and very high erosion risk in compare with other land uses. The results revealed that the erosion probabilities of rain-fed and irrigated agricultural lands were moderate and low but their erosion risks were low, considering the erosion consequences in these land uses. A hypothetical degradation scenario (modification of rangelands with 15 to 30 percent slope into rain-fed agricultural land) was used to evaluate the model performance and assesses the soil erosion risks changes in the study area. According to the results, the area with modified land use had very high erosion probability and moderate erosion risk due to the moderate values of rain-fed agricultural fields in the study area. This study highlights the importance of integrating the soil erosion probabilities of various land uses with their consequences to derive and map the soil erosion risks. This method can assists managers to prioritize the areas for various management practices based on their soil erosion risks. Keywords : Erosion, LUIM model, Bayesian belief networks, Karchambooy catchment.