The paper presents the development and validation of a mathematical model for forecasting grain crop yield based on remote sensing data and ground-based monitoring under forest-steppe conditions. The methodological framework integrates dynamic yield modeling with regression analysis of vegetation indices derived from multispectral imagery captured by unmanned aerial vehicles. NDVI-based curvilinear integrals calculated over different growth stages were used as explanatory variables in multiple linear regression models for spring wheat, spring barley, and oats. The proposed model structure incorporates trend, cyclic, and meteorological components, enabling consideration of soil and climatic factors. The regression models demonstrated high predictive performance for spring wheat and spring barley, with coefficients of determination up to 0.89 and relative errors of 17.2% and 17.5%, respectively. For oats, predictive accuracy was lower, with a relative error of 29.4%, indicating the need for further refinement. The results confirm the effectiveness of integrating NDVI dynamics and agrotechnological parameters into yield forecasting models and highlight the potential of remote sensing–based approaches for decision support and resource optimization in precision agriculture.