The conventional measuring methods (runoff plots and soil morphological comparison) and models (WaTEM/SEDEM and regional model of Russian State Hydrological Institute (SHI)) were tested with regard to the Southern Cis-Ural region of Russia, along with data from rainfall simulation for assessing soil erosion. Compared with conventional methods, which require long-running field observations, using erosion models and rainfall simulation is less time-consuming and is found to be fairly accurate for assessing long-term average rates of soil erosion and deposition. In this context, 137Cs can also be used as a marker of soil redistribution on the slope. The data of soil loss and sedimentation rates obtained by using conventional measuring methods were in agreement with the data based on the used contemporary modeling approaches. According to the erosion model calculations and data on the fallout of radionuclides in the Southern Cis-Ural (54°50–25′ N and 55°44–50′ E), the average long-term annual soil losses were ~1.3 t·ha−1 yr−1 in moderate (5°) arable slopes and ~0.2 t·ha−1 yr−1 in meadows. In forests, surface erosion is negligible, or its rates are similar to the rate of soil formation of clay–illuvial chernozems. The rates of soil erosion and sediment deposition on the arable land obtained using different methods were found to be very close. All the methods, including the WaTEM/SEDEM, allowed us to measure both soil erosion and intra-slope sedimentation. The regional SHI model fairly accurately assesses soil erosion in the years when erosion events occurred; however, soil erosion as a result of snowmelt did not occur every year, which should be taken into account when modeling. The concentrations of 137Cs in the topsoil layer (0–20 cm) varied from 0.9 to 9.8 Bq·kg−1, and the 137Cs inventories were 1.6–5.1 kBq·m−2, with the highest values found under the forest. The air dose rate in the forest was higher than in open areas and above the average of 0.12 μSv·h−1 on the slope (0.1 μSv·h−1 in the meadow and 0.08 μSv·h−1 on the arable land), with the value increasing from the watershed to the lower part of the slope in all the areas. The γ-background level in the studied ecosystems did not exceed the maximum permissible levels.