Agricultural decision support systems (ADSSs) are designed to translate complex data into actionable insights for farm management. However, a significant gap persists between the growing skill of subseasonal-to-seasonal (S2S) climate forecasts and their operational use in supporting tactical agricultural decisions. This review critically examines the current state and future trajectory of ADSSs, with a focus on bridging this gap. The analysis reveals that although modern ADSSs support a range of operational and strategic applications using historical data, predictive models, and scenario-based simulations, they largely fail to integrate operational S2S forecasts, leaving farmers vulnerable to near-term climate anomalies. Advances and challenges are synthesized across three interconnected fronts: the methodological pipeline for integrating S2S forecasts, including downscaling, bias correction, and uncertainty quantification; the imperative of participatory design and coproduction to enhance usability and adoption; and the transformative potential of artificial intelligence and machine learning under the emerging Agriculture 5.0 paradigm. The next generation of ADSSs must evolve into interactive, uncertainty-aware platforms that facilitate exploratory decision-making through bidirectional feedback loops. By synthesizing these insights, this review proposes a conceptual framework and research agenda for developing ADSSs capable of truly supporting climate-resilient agriculture.