Soil erosion and deposition in watersheds are considered environmental concerns and their control is important for the conservation of natural resources. Conventional soil practices in uplands and improper land use planning and deforestation induced by human are the main causes of soil erosion. Deforestation as one of the main components of land use change is considered as the greatest threat to global environmental diversity. When the soil plant cover is lost after deforestation, it is exposed to climatic parameters, and consequently, soil structure becomes unstable due to the decomposition of organic matter and becomes very sensitive to erosion and traortation. In order to successfully perform soil conservation programs, information about the soil loss throughout the landscape is needed. One of the well-known approaches to predict soil loss in midterm is to use 137 Cs radionuclide with a half-life of 30.2 years. In this study, the soil erosion rate was assessed in two land uses and four slope positions after the conversion of natural forest to dryland farming in Lordegan district, central Iran. A dryland hillslope and a natural forest hillslope were selected. In each hillslope, soil samples were taken from the depths of 0-10, 10-20 and 20-50 cm in four landscape positions including summit, shoulder, backslope and footslope in three transects 20 m away from each other. The activity of 137 CS was determined in all the soil samples by a gamma spectrometer. Also, the magnetic susceptibility of soil samples was measured by a Barrington Dual frequency sensor at two frequencies (0.46 and 4.6 KHz). Moreover, some soil physical and chemical properties including pH, EC, available phosphorus (P ava ), available potassium, total nitrogen (TN), organic matter (OM), gravel content, bulk density, particle size distribution and calcium carbonate equivalent were measured. The results showed that in forest land use, the highest correlation was obtained between 137 CS and magnetic susceptibility at high frequency (r=0.78, P 0.01). 137 CS showed negative and significant relationship with clay content (r= -0.399, P 0.05), while 137 CS indicated positive correlation (r=-0.345, P 0.05) with clay content in dryland farming. The results of multiple linear regression showed that magnetic susceptibility at low frequency and sand were included in the MLR developed model and could explain about 73 % of the total variability of 137 CS in forest soils. Alternatively, in dryland farming, available phosphorous, clay content and calcium carbonate equivalent were included in the models and could explain about 65% of the variability of 137 CS. 137 CS data were used to predict soil redistribution in two land uses in four landscape positions. The results of soil loss estimation by 137 CS indicated that the highest soil loss in dryland framing was observed in summit position (162.26 t ha -1 y -1 ), while the highest in forest soils was observed in shoulder position (32.35 t ha -1 y -1 ). The lowest soil loss in forest soil was observed in summit (7.2 t ha -1 y -1 ) and in backslope position of dryland farming soil (22.11 t ha -1 y -1 ). The highest soil deposition rate in two land uses were observed in footslope positions, 376 and 140.95 t ha -1 y -1 for forest and dryland, respectively. Magnetic susceptibility in two land uses reduces with depth and the highest and lowest magnetic susceptibility values were found in footslope and summit position, respectively. In forest soils, magnetic susceptibility changes significantly among the slope positions, while in the dryland no significant difference was observed among slope positions. Overal, it is concluded that deforestation has caused significant decline in soil quality indicators and 137CS technique could successfully assess the soil redistribution in the studied region. Keywords : Cesium radionuclide, Soil erosion, land use change, deforestation, dryland farming