Plant diseases cause dramatic yield losses worldwide. Current disease control practices can be deleterious for the environment and human health, calling for alternative and sustainable management regimes. Soils harbour microorganisms that can efficiently suppress pathogens. Uncovering mediators driving their functioning in the field still remains challenging, but represents an essential step in order to develop strategies for increased soil health. We set up plant communities of varying richness to experimentally test the potential of soils differing in plant community history to suppress the pathogen Rhizoctonia solani. The results indicate that plant communities shape soil-disease suppression via changes in abiotic soil properties and the abundance of bacterial groups including species of the genera Actinomyces, Bacillus and Pseudomonas. Further, the results suggest that pairwise interactions between specific plant species strongly affect soil suppressiveness. Using structural equation modelling, we provide a pathway orientated framework showing how the complex interactions between plants, soil and microorganisms jointly shape soil suppressiveness. Our results stress the importance of plant community composition as a determinant of soil functioning, such as the disease suppressive potential of soils.