Abstract: A critical need for farmers, particularly those in arid and semiarid areas is to have a reliable, accurate and reasonably accessible means of estimating the evapotranspiration rates of their crops to optimize their irrigation requirements. Evapotranspiration is a crucial process because of its influence on the precipitation that is returned to the atmosphere. The calculation of this variable often starts from the estimation of reference evapotranspiration, for which a variety of methods have been developed. However, these methods are very complex either theoretically and/or because of the large amount of parameters on which they are based, which makes the development of a simple and reliable methodology for the prediction of this variable important. This research combined three concepts such as cluster analysis, multiple linear regression (MLR), and Voronoi diagrams to achieve that end. Cluster analysis divided the study area into groups based on its weather characteristics, whose locations were then delimited by drawing the Voronoi regions associated with them. Regression equations were built to predict daily reference evapotranspiration in each cluster using basic climate variables produced in forecasts made by meteorological agencies. Finally, the Voronoi diagrams were used again to regionalize the crop coefficients and calculate evapotranspiration from the values of reference evapotranspiration derived from the regression models. These operations were applied to the Valencian region (Spain), a Mediterranean area which is partly semiarid and for which evapotranspiration is a critical issue. The results demonstrated the usefulness and accuracy of the methodology to predict the water demands of crops and hence enable farmers to plan their irrigation needs.