The Arctic has experienced much greater warming than the global average in recent decades due to polar amplification. Warming has induced ecological changes that have impacted climate carbon-cycle feedbacks, making it important to understand the climate and vegetation controls on carbon (C) dynamics. Here we used the Holocene Thermal Maximum (HTM, 11–9 ka BP, 1 ka BP = 1000 cal yr before present) in Alaska as a case study to examine how ecosystem Cdynamics responded to the past warming climate using an integrated approach of combining paleoecological reconstructions and ecosystem modeling. Our paleoecological synthesis showed expansion of deciduous broadleaf forest (dominated by Populus) into tundra and the establishment of boreal evergreen needleleaf and mixed forest during the second half of the HTM under a warmer- and wetter-than-before climate, coincident with the occurrence of the highest net primary productivity, cumulative net ecosystem productivity, soil C accumulation and CH₄ emissions. These series of ecological and biogeochemical shifts mirrored the solar insolation and subsequent temperature and precipitation patterns during HTM, indicating the importance of climate controls on C dynamics. Our simulated regional estimate of CH₄ emission rates from Alaska during the HTM ranged from 3.5 to 6.4 Tg CH₄ yr⁻¹ and highest annual NPP of 470 Tg C yr⁻¹, significantly higher than previously reported modern estimates. Our results show that the differences in static vegetation distribution maps used in simulations of different time slices have greater influence on modeled C dynamics than climatic fields within each time slice, highlighting the importance of incorporating vegetation community dynamics and their responses to climatic conditions in long-term biogeochemical modeling.