Agriculture is one of the most important branches of the national economy and the main supplier of food and raw materials for many industries. Agricultural sector in Russia has recently been undergoing renewal and growth due to the intensifi cation and application of modern innovative technologies for monitoring the state of fields using satellite images based on computer vision systems. At the same time, there is still a number of problems and challenges that require prompt solutions. One of them is developing new forecasting models and methods for key resulting indicators of agricultural development and have an advantage over existing models. To improve the accuracy of forecasting models, it is necessary to rely on a broad range of available statistical indicators and new modern econometric tools. The paper presents a set of methodological developments for modeling and forecasting crop yields based on the use of new econometric models that allow working with a truncated regression by limiting the range of possible negative values, statistical estimations of the introduced indicators that focus on the ecological component, as well as structural and general economic indicators. The suggested models allow obtaining more accurate forecasts compared to traditional popular models based on the least squares method. The work relies on Rosstat data for 100 agricultural fields located in municipalities of 43 regions of Russia, selected in proportion to the volume of crop production in this region. The results of this study are of interest to international and Russian organizations of various levels, whose activities are related to the issues of making managerial decisions aimed at ensuring food security of the country, improving the level and quality of life of the population, as well as organizations designed to provide modern conditions for farming on the ground.