The water demands of plants and soil microbes vary across seasons, yet how seasonal variations in precipitation influence soil carbon emissions remains poorly understood. We conducted a decade-long field manipulation experiment (2015-2024) in a temperate steppe in Inner Mongolia to assess the effects of seasonal precipitation changes on soil respiration (SR; total soil CO2 emissions) and heterotrophic respiration (HR; microbial decomposition-derived CO2 emissions). Our results showed that: 1) decreased precipitation in both early and late growing seasons suppressed SR (-21.1% and-20.9%) and HR (-17.5% and-16.2%), while increased precipitation in the early growing season enhanced SR (+15.9%) and HR (+15.1%), indicating negative asymmetry of SR and HR to precipitation changes; 2) during the extreme drought year (2022: 38.1% below the long-term growing-season mean (April-September)), the stimulation of SR and HR by increased precipitation during both the early and late growing seasons was stronger than the suppression caused by decreased precipitation, indicating positive asymmetry. In contrast, during the extreme wet year (2024: 39.1% above the long-term growing-season mean (April-September)), decreased precipitation suppressed SR and HR while increased precipitation had no effect, showing negative asymmetry. These findings highlighted the crucial role of initial environmental conditions in regulating soil carbon emissions; 3) respiratory responses exhibited seasonal dependence, shifting from strong associations with ANPP and microbial biomass in the early growing season to closer links with MBC in the late growing season; and 4) SR and HR increased nonlinearly along precipitation gradients, whereas the HR/SR ratio declined linearly, indicating autotrophic respiration responded more strongly than HR under enhanced soil water availability. Overall, our findings demonstrated that the responses of soil carbon emissions to precipitation changes are asymmetric and season-dependent, mediated by distinct plant (ANPP) and microbial (biomass) pathways, thereby providing mechanistic insight into carbon-climate feedbacks in semi-arid ecosystems.