Agricultural industry deals with a complex of natural factors with their self-regulation and control mechanisms. There is always a lot of uncertainty and associated risks in natural systems. Therefore, the problem of assessing these risks arises naturally, and their correct assessment is crucial in arid zones especially. The soil resource is a central component in the complex of risk factors. It ensures production stability and significantly affects the forecast of any agricultural project. However, the assessment of the soil resource itself is a complex problem, including dozens of parameters that are rely upon each other and dynamically change. The goal of the study is to solve this problem by developing the instrument for the soil resources risk assessment. There are a remarkable number of studies describing various methods and decision support systems in agriculture including soil resources assessment methods. Nevertheless, their accuracy is still insufficient. Moreover, the most accurate models consider the soil and region's climate so finely that they can barely be adapted to another conditions. This study attempts to adapt the complex assessment method to assess the risks of soil resources. This method works fine in the field of organizational and technical systems management. Its main benefit is an easy-to-construct structure allowing to simultaneously consider the variety of relationships between the evaluated parameters, including those related to the cultivated crop. The authors have modified the method to assess risks to soil resources within a given planning horizon under conditions of probabilistic uncertainty. The paper provides an example of building a risk assessment system for soil resources in arid zones.