Rewetting of dry soils usually stimulates soil carbon (C) emission, a phenomenon known as the Birch effect. Soil C cycling in drylands, which store approximately one third of terrestrial soil organic C (SOC), is strongly affected by wetting–drying cycles. However, the physical, chemical, and biological mechanisms that link rewetting cycles with dryland soil C cycling have not been comprehensively studied, nor do we understand how these mechanisms interact with each other. Here, we conducted a dryland soil incubation experiment manipulating four factors related to global change (soil moisture content, soil moisture variability, C availability, and prior warming) in a factorial design. The experiment was divided into two periods: a rewetting period consisting of six 14-d wetting–drying cycles; and a recovery period lasting 28 days during which soil moisture content was held stable, allowing for examination of the legacy effects of the wet-dry cycles. Rewetting cycles decreased soil aggregate stability under some conditions, but their effects on soil microbial biomass and fungal communities, soil enzyme activities, soil priming, and soil dissolved C were not significant. We found lower average soil respiration under the wetting–drying treatment than the stable soil moisture treatment, and Birch effects were observed, but only under some conditions. This was probably because moisture variability exacerbated soil microbial metabolic stress, which showed itself as oxygen limitation during the initial precipitation pulse and as water limitation during soil drying. Notably, respiration rates remained low even after moisture fluctuations ceased, suggesting a legacy effect of rewetting cycles on dryland microbial communities. Overall, rewetting inhibited aggregate formation (physical mechanism), and suppressed soil respiration by inducing soil microbial metabolic stress (biological mechanism), ultimately leading to lower soil C loss under rewetting. Our findings indicate that Birch effects are mediated by the magnitude of moisture variability, the availability of C, and the degree of physiological stress microbes experience.