Large storms are a major source of natural disturbance in forest ecosystems in coastal regions of the world. With the changing climate, some climate projection models predict that storms will become stronger in the future, especially in the Northwest Pacific. Although the destruction of typhoons has a strong impact on the structure and carbon balance of forest ecosystems, leading to local climate change feedback, the long-term effect of typhoon frequency and intensity on forest and carbon dynamics has not been investigated from the perspective of future climate change. To elucidate these long-term effects, we adapted a dynamic vegetation model to a deciduous-coniferous mixed forest in southern Hokkaido, Japan, to represent the development of a larch plantation established in 1954 after typhoon Toyamaru struck a natural forest and the natural regeneration from 2004 when the plantation suffered windthrow by typhoon Songda. The model was validated using eddy flux and biomass data. The adapted model was then forced with various combinations of typhoon frequency, typhoon intensity, and projected climatic conditions for 2017–2100 under RCP2.6 and RCP8.5. These future projections suggest that increasing typhoon frequency and intensity has a significant effect on forest and carbon dynamics under both RCP2.6 and RCP8.5. As the frequency and intensity of typhoons increase, aboveground biomass (AGB) and net primary production (NPP) decrease. However, increasing typhoon frequency and intensity had a greater influence on AGB than NPP. The values of AGB and NPP are much higher under RCP8.5 than under RCP2.6 and at baseline. As the frequency and intensity of typhoons increase, net ecosystem production first increases and then decreases. Our projection indicates that the forest at the Tomakomai site will become a carbon sink under different typhoon scenarios at the end of the 21st century. Our result also predicts that if typhoon frequency is ≤2 (i.e., a typhoon occurs up to two times between 2017 and 2100) and intensity is ≤40% (i.e., up to 40% of trees damaged), then forest AGB can recover to that before typhoon Songda by the end of the 21st century. Simulation results also showed strong linear relationships among forest structural traits (AGB, tree height, tree density, and woody plant and grass leaf area indexes), carbon fluxes, and typhoon frequency and intensity.