Abstract <p>The essence of the Cobb–Douglas function for agricultural production is obvious: crop yield/productivity is a function of labor and capital. At the same time, soil indicators in one form or another (humus content, plant macronutrients resulting from the application of manure and mineral fertilizers, etc.) can serve as indicators of labor and capital. The relevance of dynamic models of agricultural development based on a modified Cobb–Douglas production function of the “economic indicators—soil indicators—climatic indicators” type under conditions of changing climate and unbalanced farming systems is shown. The problems of collecting homogeneous (comparable) soil indicators over several decades are considered. They can be solved using the results of continuous agrochemical surveys (which are usually conducted by regional agrochemical services in five-year cycles), using remote sensing methods for soil and vegetation cover, getting acquainted with information contained in government reports and statistics, in soil-geographic databases (soil atlases and registers of soil resources). It is also possible to use indicators of stable soil conditions related to “soil memory”; they can characterize the degree of soil degradation, which will not change significantly during the studied period of time. The article presents databases of economic, soil, and climatic data for Belgorod and Lipetsk oblasts for the period of 1995–2017, where some soil indicators (agrochemical properties) are given in dynamics, while others (indicators of erosion) are the results of one–time measurements.</p>