The accumulation of microbial-derived organic carbon in soils during long-term vegetation succession represents a key pathway for long-term soil carbon stabilization. Yet, the role of soil microfauna in regulating microbial carbon dynamics remains poorly understood, particularly across divergent successional trajectories (primary vs. secondary successional gradients). We investigated nematode-microbe interactions influencing carbon dynamics in primary (glacial retreat) and secondary (post-disturbance) successional chronosequence on the eastern Qinghai-Tibet Plateau, to identify context-dependent mechanisms underpinning soil carbon formation and dynamics. Both successional types displayed an ‘S-shaped’ trajectory of carbon dynamics, initially rising from early to intermediate stages, declining subsequently, then increasing again in late stages, with microbial necromass (quantified by amino sugars) dominating the carbon pool, especially in later stages. In primary succession, characterized by nutrient limitation and slower soil development, nematode-mediated shifts in microbial community composition and functional gene expression, particularly reductions in growth-related and decomposition genes, enhanced microbial turnover and necromass accumulation. In contrast, in phosphorus-limited soils of secondary succession, where nutrient availability and disturbance history play a larger role, declining omnivorous nematodes modulated microbial gene expression to enhance carbon fixation and constrain decomposition under phosphorus-rich conditions. Soil pH consistently acted as the primary abiotic factor influencing microbial communities across both successional sequences, explaining up to 29% of variation in microbial gene profiles. These findings position microbial necromass accumulation as a central pathway for carbon sequestration and identify nematodes as context-dependent regulators of microbial carbon processing. Our study underscores the need to integrate both soil microbial and faunal dynamics, along with key physicochemical properties, into ecosystem models to improve predictions of carbon-climate feedbacks across successional landscapes.