Agriculture accounts for nearly 70% of global freshwater withdrawals, making efficient irrigation management critical for sustainability. Traditional irrigation practices often rely on fixed schedules or farmer intuition, which can result in excessive water use and reduced crop productivity. This paper presents an artificial intelligence (AI)-based framework for water optimisation in agriculture through predictive modeling and decision support. The approach integrates real-time data from soil moisture sensors, weather forecasts, and crop indices such as the Normalized Difference Vegetation Index (NDVI) to predict short-term water requirements. Machine learning models, including regression and time-series forecasting techniques, are used to estimate soil moisture and crop evapotranspiration. These predictions feed into an optimisation algorithm that schedules irrigation with the objective of minimizing water usage while maintaining crop yield. Experimental trials demonstrate that the proposed system can reduce irrigation water by approximately 20–30% compared to conventional methods, without significant yield reduction. The findings highlight the potential of AI-driven irrigation to improve water use efficiency, reduce input costs, and promote sustainable farming practices.