Background and aims <p>Enhancing soil carbon sequestration requires a thorough understanding of the interactions among plants, rhizosphere-associated microbiota and soil properties. Here, we investigated the interplay between pearl millet (<i>Pennisetum glaucum</i>) genotypes and land use on carbon sequestration within the rhizosheath, as assessed by net carbon gain (NCG), and on the structure of rhizosheath-associated microbiota.</p> Methods <p>We used the same arenosol from adjacent forest and vineyard sites. Two pearl millet lines with contrasting rhizosheath sizes assessed genotype effects. Microbial diversity was analyzed via 16S and ITS2 rDNA metabarcoding for bacteria and fungi, respectively.</p> Results <p>Our findings revealed a significant interaction between land use and plant genotype, resulting in a greater rhizosheath size for the high-aggregation line, compared to the low-aggregation line, but only in forest soil. This interaction was also significant for the NCG: the land use effect (NCG_forest &gt; NCG_vineyard) was significant only for the high-aggregation line. We founded that the bacterial reservoir was significantly different between the land uses. In root tissue, bacterial taxa such as <i>Burkholderia</i>, <i>Rhizobium</i> sensu lato, and <i>Sphingomonas</i>—known for their exopolysaccharide production—were consistently more prevalent than in the rhizosheath and bulk soil. Additionally, predicted abundances of genes such as <i>exoP</i>/<i>vpsO</i>, encoding an exopolysaccharide (EPS) biosynthesis transport protein, were significantly higher in forest soil.</p> Conclusions <p>Rhizosheath formation depends on two main factors: (i) microbial diversity shaped by land use, including exopolysaccharide-producing bacteria, and (ii) rhizodeposit composition and carbon input influenced by plant genotype.</p>