Wood logging is one of the most common anthropogenic disturbances reshaping the historical and future global forest patterns and contributing a large amount to land use emissions. However, the impacts of wood logging on land biophysical properties, when also taking into account ecosystem demography, canopy structure, and biogeochemical cycling lack comprehensive global studies. Terrestrial ecosystem models are common tools to simulate wood logging and its consequence on terrestrial carbon cycle. Global land use drivers for these models often use estimates of harvested area as their forcing. Carbon-based harvest data are usually available from countries’ historical forest inventory or regional censuses, providing more precise wood harvested carbon compared to area-based data. We explore the consequence of wood logging by incorporating a global C-based harvest algorithm into the Functionally Assembled Terrestrial Ecosystem Simulator (FATES), a vegetation demographic model with full ability to simulate dynamic vegetation and land cover heterogeneity dynamics. We then drive the model using both the area-based and carbon-based harvest estimates from the Land Use Harmonized Dataset v2 (LUH2) dataset. FATES separately resolves the dynamics of both primary and secondary lands, with separate land use drivers on each. We evaluate the harvested carbon and land area for both primary and secondary lands in FATES with the LUH2 products. We compare three historical experiments from 1850 to the recent decade: 1) no global wood logging, versus 2) historical wood logging with harvest driven by LUH2 harvest area, and 3) historical wood logging with harvest driven by LUH2 harvest carbon, to estimate the potential global impact on modeled land physical properties, vegetation carbon storage, and LUC induced emissions from both primary and secondary lands. The results show an increasing carbon sink from the contribution of secondary forest land for recent decades. By comparing to the carbon-driven and area-driven harvest amounts, we find that the carbon-based logging reduces the overestimates of losses in ecosystem carbon storage over regions with denser forest, and low biases over sparser forest area.