This paper proposes an empirical climate classification method based on the application of multivariate statistics. In this method, the technique of supervised and unsupervised image classification is used to classify the data and define climatic units. envi software is used to perform the image classification, specifically, the Iterative Self‐Organizing Data Analysis Technique algorithm. Supervised classification is also applied based on reference variables, fundamental parameters and a classifier. The obtained results display greater objectivity, reliability, operability, accessibility and reproducibility than previous climate classifications devised for the region of Andalusia (Spain), taking into account that these previous classifications were not based on quantitative criteria.