Reference evapotranspiration (ETo) is a key parameter for efficient irrigation management, especially in protected agriculture systems located in arid and semi-arid climates. This study aimed to develop a predictive ETo model within a shade net greenhouse using evolutionary computing techniques, specifically genetic programming (GP). For model development, climatic data were collected at one-minute intervals, allowing for high-resolution capture of the specific microclimatic conditions within the protected environment. The model was trained using ETo values calculated with the FAO-recommended Penman-Monteith method (FAO56-PM), considered the international standard for reference evapotranspiration estimation. The resulting model exhibited outstanding performance, achieving a root mean square error (RMSE) of 0.217 and a coefficient of determination (R-2) of 0.99, indicating high predictive accuracy. When compared to widely used empirical models such as Hargreaves-Samani (RMSE = 4.82, R-2 = -0.98) and Priestley-Taylor (RMSE = 1.00, R-2 = 0.913), the proposed model significantly outperformed both traditional approaches. These results highlight the potential of genetic programming as an effective tool for developing robust predictive models tailored to specific conditions, such as those found in protected agricultural systems under arid and semi-arid climates.