Abstract: Smart agriculture is driving the digital, intelligent, and intensive transformation of agriculture in China, creating demand for more flexible monitoring and field operations. As part of integrated aerial, satellite, and ground-based systems, agricultural low-altitude technology can extend monitoring coverage, improve operational efficiency, and support operations in complex terrain where conventional ground equipment is constrained. This review takes agricultural unmanned aerial vehicles (UAVs) as its principal platform. It synthesizes research from China and other countries and establishes a framework linking low-altitude sensing, intelligent decision-making, and precision execution. At the sensing stage, red-green-blue, multispectral, hyperspectral, thermal infrared, light detection and ranging, and micrometeorological systems acquire information on crops, soils, terrain, and field conditions. Advances in deep learning and multisource data fusion support the detection and localization of weeds, diseases, lodging, and seedlings. These methods also enable assessment of crop growth, canopy structure, nutrient and water status, and yield. At the decision stage, sensed information is translated into operational prescriptions and control parameters. Autonomous flight has progressed from fixed routes toward terrain-adaptive planning, obstacle avoidance, energy-aware optimization, and dynamic replanning. Multi-UAV cooperation increasingly involves dynamic task allocation and coordination with ground equipment. Crop models, artificial intelligence, digital twins, and edge–cloud computing further support prescription generation and operational scheduling. At the execution stage, precision plant protection is advancing from uniform spraying toward variable-rate application, site-specific treatment, and targeted release of biological control agents. Precision sowing and fertilizer spreading are incorporating variable discharge, material-specific control, online flow monitoring, and deposition prediction. Specialized applications, including agricultural-product lifting and transport, UAV-assisted pollination, and crop detasseling, extend UAVs beyond general-purpose spraying platforms. Together, these developments indicate a transition from separate sensing and aerial application activities toward coordinated systems connecting field observations, management decisions, and precise operations. This information-driven framework allows monitoring results to guide flight plans and variable-rate interventions. However, wider deployment remains constrained by limited endurance and payload, inconsistent sensing performance in variable field conditions, and decision algorithms that transfer poorly across crops and regions. Incomplete technical standards, uneven adaptation to fragmented farmland and complex terrain, insufficient service systems, and differing airspace-management requirements further impede routine use. High purchase and maintenance costs, a shortage of trained personnel, and limited support in remote rural areas also restrict adoption among smaller producers. Future development should prioritize higher-performance autonomous equipment, robust multisource sensing, adaptable decision models, and coordinated aerial and ground operations. Standardized evaluation and application protocols should be tailored to specific tasks and production scenarios. Service networks, digital management platforms, policy support, and category-based airspace governance could strengthen large-scale implementation and operational oversight. Together, these measures could advance autonomous equipment, integrated perception, intelligent decisions, collaborative operations, and standardized management, supporting the systematic development and sustained use of agricultural low-altitude technology in smart agriculture.