The Qinghai-Tibet Plateau has undergone significant anthropogenic activities and climate change impacts over the past years, with these trends projected to intensify. These changes are likely to alter the alpine grassland ecosystem structure and services. To assess the impacts on carbon storage (CS), we employed an analytical framework integrating the optimal parameter-based geographical detector (GD) model, system dynamics (SD) model, patch-generating land use simulation (PLUS) model, and Integrated Valuation of Ecosystem Service and Tradeoffs (InVEST) model. Applying this framework to the Qinghai Plateau, we used the GD model to elucidate the spatiotemporal evolution patterns and driving mechanisms of changes in grassland types for alpine grassland ecosystem. The SD and PLUS models simulated the spatial distribution of grassland types under the coupled scenarios of three typical shared socioeconomic pathways (SSPs) and representative concentration pathways (RCPs) (SSP-RPCs: SSP126, SSP245, and SSP585). InVEST was used to evaluated the impacts on CS. Our findings revealed that: (1) From 1990 to 2020, grasslands on the Qinghai Plateau expanded by 9.7 %, with significant shifts towards types more suitable for humid environments, including a transformation of 2.1 × 104 km2 of of alpine steppes into alpine meadows; (2) Changes in grassland types are driven by both anthropogenic activities and natural environmental factors, notably climate change and grazing intensity; (3) Over the past three decades, grassland CS increased by 6.687 × 108 t, with expansion of grassland area and shifts in grassland types contributing 68.2 % and 31.8 %, respectively; (4) The integration of the PLUS and SD models enabled simulations of grassland-type distribution, achieving an overall accuracy (OA) of 0.86 and a Kappa statistic of 0.78. (5) Under SSP-RPCs development scenarios, grassland and alpine meadow areas are projected to continue expanding due to ongoing warming and humidification. The most significant expansion and the largest increase in CS were observed under the SSP585 scenario. These results are crucial for understanding the evolutionary patterns of alpine grassland ecosystem and their impact on regional carbon balance, offering valuable insights for ecosystem management and conservation strategies.