Tackling the Sicilian woody vegetation as a case-study, this work aims to verify the correspondence between Rivas-Martínez's bioclimatic units and the main vegetation series in the Mediterranean region. Following this approach, one macrobioclimate and 25 bioclimatic type belts can be recognized in Sicily. By means of a geostatistical analysis based on WorldClim data sets, cartographic models of the distribution range of each single bioclimatic unit were obtained and combined with vegetation data, in order to develop a new regional spatial framework, integrating climatic and vegetation data. Fidelity of each vegetation unit to a given climatic range was then evaluated as percent distribution of the occupied surface within a given bioclimatic unit, while the predictive power of the WorldClim data sets was tested by using half of the spatial data of the processed vegetation units as independent variables. Our results suggest that: (1) any kind of numerical threshold used to define bioclimatic units is not effective a priori, but only after it has been adjusted to the territory and to the spatial scale used to set the model; (2) bioclimatic indices being an empirical tool, the model can be trained and eventually adjusted when applied to different territories; (3) fidelity of vegetation units to a given bioclimatic unit is highly variable; (4) the mechanistic pitfall that climatophilous vegetation has to be necessarily linked to a single bioclimatic unit should be avoided.