Forestation plays a pivotal role in mitigating climate change by capturing a substantial fraction of CO2 emitted by activities. However, the efficiency of carbon (C) sequestration for different restoration approaches such as planted forests (PFs) and naturally regenerated forests (NRFs) remains unclear, causing large uncertainty in climate projections. A terrestrial ecosystem model framework, which combined the Bayesian approach with a traceability method, was used to identify the efficacy of C sequestration in four NRFs and two PFs after 54-years of reforestation in the boreal forest of northeastern China. The results showed that the rate of C accumulation among the six forests ranged from 135.1 (Mongolian oak) to 400.3 g C m-2 year-1 (Dahurian larch) in the period of 2015-2100. The two highest total C storage values at the end of the projection period were Dahurian larch forest (57.9 ± 1.59 kg C m-2, PFs) and hardwood forest (53.8 ± 3.01 kg C m-2, NRFs), which were mainly the result of net primary production (NPP, 749.1 ± 47.2 g C m-2 year-1) and ecosystem C residence time (ζ_E, 101 ± 5 years), respectively. The efficiency of C sequestration in PFs was found to be more influenced by plant traits, especially NPP, as well as the turnover rate of foliage and carbon allocation to wood. In contrast, the efficiency of C sequestration in NRFs was more correlated to the residence time of soil organic carbon, the C:N ratio of litter, and the content of soil inorganic nitrogen. The study highlights the different contributions of plants and soil to carbon storage in PFs and NRFs and advances our understanding of C accumulation between PFs and NRFs forests in long term under the same climate condition. Our findings provide valuable guidance for designing reforestation strategies aimed at enhancing the terrestrial carbon sink in the future climates.