Land use variations influence aquatic dissolved organic matter (DOM); however, the responses of relationships between DOM composition and microbial communities/metabolic genes to land use types and landscape indices remain insufficiently explored. This study conducted four field sampling sessions on the Qingshui River, northwestern China, investigated the shifts in DOM-transforming microbes and genes under distinct land use patterns (forest & grassland, urban area, and agricultural land) using metagenomics analysis, elucidated the DOM-microbe interactions in response to land use characteristics from the perspective of both basin and land use types by statistical approaches. At the basin scale, land use primarily influenced DOM quality rather than quantity. Partial least-squares path model demonstrated that landscape indices influenced DOM components predominantly by inhibiting gene abundances associated with metabolic pathways, exerting a more pronounced impact than land use area proportions. Notably, DOM composition, microbial metabolic genes, and their interactions differed significantly across land use types. Urban areas exhibited the highest levels of DOM aromaticity and the largest proportions of autochthonous sources, their fragmented landscape characteristics destabilized the microbial network, resulting in a comprehensive rather than partial suppression of genes involved in DOM transformation pathways (carbohydrate metabolism, energy metabolism, and amino acid metabolism). Compared to other land use scenarios, the downstream agricultural areas emerged as active zones for microbial utilization of DOM, maintained high microbial network stability and gene abundances similar to forest & grassland, highlighting the microbial resilience. These insights advance mechanistic understanding of how land use governs aquatic DOM-microbe interactions in arid and semi-arid watersheds.