Background and AimsPerennial crops often improve soil structure relative to annual systems, but the biochemical pathways by which plants influence soil aggregation remain unclear. We tested whether seven years of Miscanthus × giganteus cultivation improved soil aggregation relative to continuous maize and whether these changes were associated with microbial- versus plant-derived biochemical compounds.MethodsWe sampled soils from three Iowa sites under maize, miscanthus, and adjacent turfgrass alleyways as a perennial reference. We measured water-stable aggregates, mean weight diameter, microbial biomass, extracellular polysaccharides, amino compounds, and extractable carbohydrates, including the microbial-to-plant carbohydrate monomer ratio (GM:AX).ResultsMiscanthus increased water-stable aggregates and mean weight diameter by ~41% and ~74%, respectively, relative to maize across sites, while turfgrass showed smaller increases (36% and 61%). Contrary to our second hypothesis, microbial-derived pools, including microbial biomass, extracellular polysaccharides, and amino compounds, showed limited differences among treatments. Instead, miscanthus uniquely altered extractable carbohydrate composition, decreasing GM:AX by ~11% relative to maize and turfgrass, indicating a stronger contribution of plant-derived carbohydrates to the labile carbon pool. In contrast, turfgrass increased aggregation alongside higher microbial biomass and amino compounds, consistent with a more microbially mediated pathway.ConclusionThese results show that perennial plants can improve soil aggregation through distinct biochemical pathways at the plant–soil interface. Miscanthus was associated primarily with shifts in plant-derived carbohydrates, whereas turfgrass showed stronger microbial signatures. This work advances mechanistic understanding of how perennial vegetation shapes soil structure and may help guide the design of cropping systems that improve soil function.