In the context of highly pressured agricultural production using low levels of inputs, cropping strategies could take advantage of plant-microbiome interactions that can improve both plant growth and health. To do so, a better comprehension of the mechanisms that drives these interactions, such as the exudation of signalling molecules by plants, is needed. However, collection and analysis of the root exudates show significant technical challenges. Indeed, root exudates are mostly studied in artificial conditions, such as in hydroponics, while they are affected by microbial presence and soil properties. Here, we compared root exudates and bacterial diversity for two tomato genotypes under three different growing conditions (i.e. in aeroponics and two natural soils). We analysed root exudates using a metabolomics approach and we assessed bacterial diversity and composition using a metabarcoding approach (16S amplicon sequencing). Using multivariate analysis, we observed a significant, mainly quantitative, effect of the growing conditions on the root exudate composition. The bacterial diversity decreased from bulk soil and the rhizosphere to the root itself. In addition, even at the phylum level, the belowground bacterial community composition was significantly different according to the compartment and the soil. However, the root bacterial diversity and composition of the two tomato genotypes were similar in aeroponics and in the two soils. This suggests that the root microbiome is not so much affected by horizontal transfer of microbes from the soil. Thus, within this study, we developed a method to analyse root exudates in soil. Further advanced statistical analyses will be used to link the root exudate profile to the microbiome composition in order to find the molecular mechanisms involved in microbiome selection by plants.