Introduction Mediterranean soils are experiencing increasingly negative effects of climate change due to drought and extreme weather phenomena that negatively affect agriculture, already suffering social and economic constraints. The soil microbiota is recognized as key player in both diversity and productivity of terrestrial ecosystems and it is mandatory to include its role in sustainable agronomic management. The LIFE Project DESERT-ADAPT - Preparing desertification areas for increased climate change - (http://www.desert-adapt.it/) focuses on agricultural adaptation measures aimed at reverting ongoing desertification trends (while improving the socioeconomic conditions of farmers). In order to improve our understanding of soil microbes and their contribution to ecosystem functioning, we analyzed the bacterial communities of soils under strong desertification risk at LIFE Desert-Adapt partners’ farms in Southern Europe. The bacterial diversity has been related to soil physical-chemical features and current soil use. Materials and methods Total DNA was extracted from 70 agricultural and agro-forest soil samples from ten farms in Spain, Portugal and Italy and used as a template in PCR amplifications for automated ribosomal intergenic spacer analysis (ARISA) using universal primers targeting the 16S-23S rRNA ITS. The output of ARISA was analyzed by the Peak Scanner software vs1.0 and the biodiversity indices where calculated using the PAST 3.0 software (http://folk.uio.no/ohammer/past). Bacterial diversity was related to soil pH, soil biomass, total organic C (TOC), N, and Cation Exchange Capacity. Results Using total soil DNA as a proxy of soil microbial biomass we found that pine tree forests and pastures have higher biomass content compared to arable and other soils. Bacterial richness, analyzed by ARISA, is higher in the Italian soils that are generally characterized by higher pH (from 7.6 to 8.6) in comparison to Spanish and Portugal soils (pH from 4.4 to 6.2). Nonparametric tests indicate that Soil OM is positively related to biomass content (p <.0001) and C/N ratio (p <.0001) and inversely proportional to bacterial richness (p=0.0439). Canonical analysis of principal coordinates (CAP) of ARISA profiles shows that pH, TOC, and soil use exert the strongest influence on the community structure. Conclusion We have chosen to focus on bacterial diversity as indicator of soil health and functionality of forest and agricultural soils threatened by desertification in Southern Europe. In the next four years the 10 landowners, partner of the LIFE Desert-Adapt Project, will implement their model of adaptation to desertification in their lands. The results of this investigation will be then used as baseline for the evaluation of the effects of the adaptation measures implemented in the ten farms at the end of the LIFE Desert-Adapt Project.