Efficient water management in agriculture is crucial for sustainable crop production, particularly in regions facing water scarcity. This article introduces a comprehensive predictive model for optimizing the current irrigation of Dutch roses in the Beni Mellal region of Morocco. The model addressed the need for precise water management across four distinct plant growth stages. The integrated system proved able to estimate the daily irrigation requirements based on historical weather data and crop-specific factors. The model incorporated four main components: weather prediction for temperature, net radiation, wind speed, and dew point; calculating the reference evapotranspiration using the Penman-Monteith equation; applying the crop coefficients specific to each growth stage; as well as estimating the crop evapotranspiration and determining daily water needs. The system offered a systematic approach to predicting the daily water requirements for Dutch roses across the entire growth cycle. By leveraging historical weather patterns and growth stage-specific crop coefficients, the system provided a predictive tool for proactive irrigation management. The model proved highly adaptable as it was able to generate forecasts based on weather trends and plant growth stages, potentially leading to a more efficient water use than conventional irrigation methods. This integrated approach is expected to allow the rose farmers of Beni Mellal to optimize their irrigation practices. While field validation is needed to quantify its impact, the model’s framework already shows potential for enhancing water use efficiency in cultivating roses and other crops in arid environment.