As global climate change intensifies, the frequency and severity of crop abiotic stress events (e.g., droughts, heatwaves, floods, and salinity) are increasing, posing severe threats to agricultural production and national food security. In China, crop abiotic stresses exert pronounced impacts on staple crops, highlighting the urgent need for accurate, timely, and scalable monitoring approaches. Traditional approaches based on field surveys, empirical indices, and expert judgment struggle to capture early-stage physiological responses and spatiotemporal dynamics, limiting their ability to meet the practical demands of intelligent and large-scale crop stress monitoring. Consequently, they fall short of the requirements of modern smart agriculture and disaster risk management. In recent years, the deep integration of remote sensing and artificial intelligence (AI) technologies, known as “AI + remote sensing”, has made significant progress in crop abiotic stress monitoring. This review provides a comprehensive overview of the integrated “space–air–ground” monitoring framework across different application scenarios. We discuss the applications of satellite platforms, unmanned aerial vehicle platforms, and ground-based sensor networks in monitoring crop abiotic stresses. The integration of these platforms provides multi-scale and multi-source datasets for crop stress detection, damage assessment, and phenotyping. Moreover, we outline the evolution from machine learning to deep learning approaches, as well as recent advances in multimodal remote sensing data fusion and explainable AI. AI-driven multimodal data fusion strategies integrate optical, thermal infrared, hyperspectral, and synthetic aperture radar data, substantially enhancing the capability for continuous, reliable, and large-scale monitoring of crop physiological responses under complex environmental conditions. In addition, emerging explainable AI approaches are highlighted for their ability to improve model interpretability and robustness while strengthening the link between model outputs and underlying agronomic or physiological mechanisms. Furthermore, we analyze the industrial demands and practical challenges of agricultural disaster prevention and mitigation in China, including data fragmentation, limited model generalization, and gaps between research outputs and operational deployment. We also explore how to address these challenges by coordinating data infrastructure, algorithmic innovation, and application-oriented system design. Overall, this review offers valuable insights into the development of intelligent agricultural disaster monitoring and early-warning systems and facilitates the translation of advanced “AI + remote sensing” technologies into practical applications for smart agriculture, sustainable crop production, and food security. By systematically reviewing technological breakthroughs in integrating remote sensing and AI technologies for monitoring crop abiotic stresses, this review provides scientific guidance for establishing smart monitoring and early-warning systems in China. It will drive the transformation of state-of-the-art research outcomes into practical technologies and operational tools across the crop production lifecycle.
Song et al. (Fri,) studied this question.