Abstract
Intelligent irrigation is increasingly positioned as a means of reconciling agricultural productivity with growing constraints on water availability, yet its effectiveness depends on how heterogeneous field observations are transformed into reliable and context-specific irrigation decisions. This review addresses the central question of how sensing, communication, analytics, decision-making, and control can be integrated into a coherent intelligent irrigation framework. We synthesize recent advances in multi-source sensing, Internet of Things (IoT) networks, data analytics and artificial intelligence (AI), automated control, and human–machine interaction, and examine their roles across the irrigation decision–control chain. Particular attention is given to the representation of spatial and temporal heterogeneity, data quality and uncertainty, model generalization, and the compatibility of intelligent decisions with hydraulic and agronomic constraints. Evidence from diverse application contexts is further assessed in terms of water productivity, crop performance, resource use, economic feasibility, and environmental outcomes, while technological, socioeconomic, and institutional barriers to large-scale deployment are critically examined. On this basis, we propose a conceptual progression from observation and heterogeneity representation to adaptive decision-making and infrastructure-enabled control, highlighting multi-source data fusion, hybrid process–AI modeling, edge–cloud collaboration, digital twins, and context-specific adaptive control as key directions for advancing system intelligence. The review reframes intelligent irrigation from a collection of digital technologies into an integrated observation–decision–control framework, providing a basis for evaluating not only what intelligent irrigation can achieve, but also where, when, and under what conditions it can deliver reliable and scalable water-management outcomes.
