Formalizing the connections existing between socio-ecosystem components implies establishing the dynamical links between both societies and vegetation cover types along time and space. In order to synthesize knowledge of Mediterranean paleo-environments and populations, the MISTRALS PaleoMex program aims at developing a step by step multidisciplinary and spatialized model of the climate, vegetation and human implantations in the Mediterranean seashores of Occitania Province (France) during the Neolithic Eras. The first step is to reconstruct the vegetation covers in the absence of humans. For this, 95 tree and bush species known to have grown in the region in Neolithic times are considered, which combining present-time species, historical data and pollen sequences. These species are then grouped in Plant Functional Types (PFTs) according to their tolerance to three factors: the mean annual temperature, the pH and the soil water balance. Two clustering methods were first tested: HCA (Hierarchical Clustering Analysis) and k-means based on the species' tolerance interval for each factors. The resulting PFTs were well-defined enough to statistically explain the total ecological variance but were misleading botanically speaking, by failing in identifying clearly well-known PFTs such as the pioneer groups. A third method was thus assessed based on group species' overlap of their tolerance intervals. Only 80% of the total variance was explained but the resulting 8 PFTs are more representative of natural species associations including a well-identified pioneer species PFT. Further investigations may be pursued to reach a total of 11 PFT groups in order to explain 95% of the total local variance. This PFT stock will be assessed using a virtual spatialized cellular automaton model with a 1ha spatial resolution and seasonal timescale. Elevation, pedology and dynamic climatology for each season and hectare will allow deciphering the spatial and dynamic evolution of the vegetation cover as a localized repartition of these PFTs.