Abstract
Crop rotation optimization has evolved from experience-based management toward a complex decision science problem involving multi-scale processes, competing objectives, and dynamic constraints under global climate change and increasing resource pressures. The rapid advancement of Big Data, remote sensing, Internet of Things, and artificial intelligence technologies has reshaped the theoretical and methodological framework of crop rotation research by enabling integrated information perception, process simulation, predictive analysis, and intelligent decision-making. This review systematically synthesizes recent advances in crop rotation optimization from the perspective of the “data–model–decision” framework. At the data level, multi-source heterogeneous observations, including remote sensing, sensor networks, and field monitoring, provide critical support for characterizing spatial–temporal variability and system states. At the modeling level, process-based models, data-driven approaches, and hybrid modeling paradigms offer complementary capabilities for mechanism interpretation, predictive accuracy, and system representation. At the decision level, multi-objective optimization, reinforcement learning, and scenario-based analysis facilitate the coordination of production efficiency, economic benefits, soil health, resource conservation, and environmental sustainability. However, key challenges remain in integrating heterogeneous data streams, balancing model interpretability and generalization, quantifying uncertainty, and translating computational solutions into operational management strategies across spatial scales. By critically comparing existing methodologies and identifying their applicability boundaries and synergistic pathways, this review proposes an integrated conceptual framework for intelligent crop rotation optimization that links observation, mechanism understanding, and adaptive decision-making. The findings provide theoretical foundations and methodological guidance for developing trustworthy, scalable, and sustainable crop rotation decision-support systems in future agricultural landscapes.
