Under conditions of changing climate, unsustainable use of agricultural land and increasing anthropogenic pressure on the environment, the natural bioproductivity of the soil is being lost. The integration of geographic information systems (GIS) and remote sensing has been shown to promote precision farming technology and sustainable land use practices. In this study, based on the Revised Universal Soil Loss Equation (RUSLE) method, a digital elevation model of an experimental site in eastern Kazakhstan was developed. It was found that one of the main causes of erosion is critically low content of humus in the soil in the range of 0.82-1.66 %, so the use of spatial modelling methods can help determine the best management practices to reduce and prevent soil erosion and improve its quality by increasing the content of humus in the soil. Calculations of actual soil losses after gully formation, which amount to about 185 ha of land, have been made; a predictive model of the relief of the study area and increase of humus content to 3.5-5% with the proposed fertility restoration measures has been developed to prevent further soil erosion, preserve and restore fertility.