Rangeland is a natural ecosystem which includes major resources of geneti reserves and different types of plant species and this biodiversity indicated stability of rangeland against environmental and biological variable factors. Soil and plant serves as an important rangeland ecosystem component and evaluating the effects of management practices on structure and function of range ecosystems can assist managers to select the best option in their reclamation plans. This study aimed to compare the effects of long term exclosure and controlled grazing (rest- rotation grazing system) on some soil physicochemical properties in semi- arid rangelands of Hanna - Semirom in Isfahan province. Six perpendicular transects were established in each area and 10 square plots with size of 2m 2 were placed along each transect and the cover percentage, production, density and list of all plant species were recorded. The area under rest rotation grazing systems and exclosure area, randomly 72 soil samples were collected from 0 - 10 and 10 - 30 cm of soil surface. Soil properties such as pH, EC, Caco 3 , organic carbon, absorbable phosphor, total nitrogen, k, Ca, Mg, soil saturated percentage, cation exchange capacity, soil clay, silt, sand and fine sand contents were measured in the soil laboratory. Non parametric diversity, richness and evenness indices and parametric abundance distribution models were calculated for both areas using PASTsoftware. Independent t test was used to compare vegetation and soil properties in both areas. Controlled grazing management cause increased of significant production and vegetation cover percentage compared to exclusion area. The vegetation density was increased significantly in exclusion compared to controlled grazing area. According to the results, although “lightly grazed” site had higher diversity compared to “not grazed” site, only Hill diversity indices varied significantly ( p 0.05) between these sites. Results indicated that richness indices varied significantly between the two sites ( p 0.05). Abundance distribution model showed that no parametric model fitted to the diversity data in “not grazed” site, indicating an unstable vegetation community. Broken stick and log- normal were best fitted parametric models among the species abundance models for the “slightly grazed” site, indicating a stable vegetation community in this site. Results revealed that soil properties were more appropriate in areas under long term rest-rotation grazing system compared with excloure area. According to the results, most soil properties were differed significantly (? = 5%) due to grazing management in both areas, except pH, Mg, soil saturated percentage, silt, sand and fine sand contents. Cation exchange capacity, CaCO 3 , gravel percentage, soil phosphor content and grazing management were identified as the most discriminative factors in separatingvegetation communities based onCanonical correspondence analysis (CCA) and cluster analysis. This study showed that rest- rotational grazing systems could maintain ecosystem health condition in semi-arid area as well as decrease the costs of animal husbandry and is a more appropriate management option for ranchers compared with long term grazing exclusion. Key words: Canonical correspondence analysis, Species diversity, Controlled grazing, Grazing system, Abundance distribution model