Groundwater-dependent irrigated agriculture in semi-arid basins is increasingly threatened by climate variability and aquifer depletion. Karapınar district, located within Türkiye’s Konya Closed Basin, represents a water-stressed melon-producing area where deep-well drip irrigation and rigid calendar-based scheduling remain common. This study developed a prototype AquaCrop-based decision-support framework for adapting mature and pickled melon irrigation to climatic conditions classified by the crop-season weighted three-month Standardized Precipitation Index (SPI-3). SPI-3 was selected to represent short-term seasonal precipitation anomalies relevant to agricultural water availability, and three historical years were used as representative scenarios: Dry (2020), Normal (2018), and Wet (2017). Fourteen irrigation treatments, including farmer practice and phenologically targeted Late-Season Event (L.S.E.) deficit strategies, were simulated using FAO AquaCrop v7.1. Irrigation timing was more important than seasonal irrigation volume alone. Under Normal conditions, an L.S.E. strategy that redistributed irrigation toward four late fruit-development events (T4.2) maintained yield stability comparable to farmer practice (CV < 2%), increased mature melon yield by 12.7% relative to farmer practice, and reduced applied irrigation by 19–21%. Under Dry conditions, a lower-volume L.S.E. strategy with the same late-season targeting (T4.3) maintained baseline-level yield while reducing applied irrigation by 28–31%. The resulting SPI-based decision matrix is forecast-compatible but not operationally validated; field validation, long-term climate testing, and forecast-skill evaluation remain necessary before deployment.