The storage of soil organic carbon (SOC) within grassland ecosystems significantly contributes to global carbon sequestration. Understanding long-term trends of SOC in grasslands is imperative for promoting climate change adaptation and mitigation endeavors. However, accurately forecasting how climate change will impact SOC dynamics remains challenging. Using the DNDC (Denitrification and Decomposition) model, we carried out simulations in this research to determine the reaction of SOC to climate change in the grasslands of Northern Xinjiang, considering four different scenarios from the phase 6 of the Coupled Model Intercomparison Project (CMIP6). The simulation revealed that the dynamics of SOC varied considerably across the scenarios. Specifically, under the SSP370 scenario, SOC was predicted to decrease by 9.78 +/- 0.74 Tg C, while under the SSP126, SSP245, and SSP585 scenarios, SOC was predicted to increase by 2.42 +/- 0.32, 1.03 +/- 0.25, and 13.40 +/- 0.94 Tg C, respectively. Furthermore, the spatial distribution of the primary climate factor causing SOC variations exhibited significant variation. According to our projections, there will be significant variation in SOC changes among different types of grassland under future climate scenarios. Alpine meadow and alpine steppe are expected to experience greater carbon loss compared to other types of grasslands under future climate scenarios. Consequently, the capacity of alpine grasslands to sequester carbon is projected to decrease. This underscores the necessity of enhancing the ability of alpine grasslands to sequester carbon, given that the response of grassland ecosystem may further accelerate the rate of climate change.