Terrestrial ecosystems play a vital role in the global carbon cycle, where vegetation carbon sequestration serves as a critical metric for assessing ecosystem health and climate regulation capacity. As China's national strategic economic hub, the Beijing-Tianjin-Hebei region holds pivotal significance in advancing regional low-carbon development and realizing the "Dual Carbon" objectives. Current carbon sink inversion methodologies encounter limitations, including weak robustness in empirical parameters (e.g., light use efficiency), resulting in substantial uncertainties in carbon sink quantification. This study uses long-term multi-source active and passive satellite remote sensing data from 2003 to 2022 to estimate the spatiotemporal changes of carbon sinks and their driving mechanisms in the region. The key steps include: Firstly, we enhanced the dynamic correction method for maximum light use efficiency through integration with GEDI LiDAR data, thereby optimizing the CASA model's net primary productivity (NPP) simulation capability; Secondly, based on the soil heterotrophic respiration model, the NEP of Beijing-Tianjin-Hebei region in the past 20 years was estimated; Finally, the Shapley Additive exPlans model was used to quantify the driving effects and nonlinear interactions of environmental factors such as temperature, precipitation, and solar radiation on vegetation carbon sinks. The results show that: (1) Based on GEDI, dynamic correction of long-term series of maximum light use efficiency has been achieved. The simulated maximum light use efficiency of vegetation has passed the accuracy validation, with the forest light use efficiency attaining peak values of 0.667–0.712 gC·MJ?1. This improved method can reliably simulate the NPP, with results indicating that the average annual NPP first decreases and then increases in the region. (2) The average annual Rh is 233.45 gC·m?2·a?1 in the region, with a spatial distribution that follows the latitudinal zonation pattern, and approximately 82% of the region shows an increasing trend in Rh. Furthermore, the average annual NEP is 179.17gC·m-2·a-1, with a seasonal variation of summer>spring>autumn>winter. Spatial gradients exhibit a decline from northern Hebei's mountainous zones to southeastern plains and urban clusters, and the contribution rate of forests to NEP is 47.60%. (3) The importance and main effect degree of each driving factor on NEP is in the order of temperature>precipitation>solar radiation. The positive interaction between temperature and precipitation is the strongest, with an interaction value of 1.587, while temperature and solar radiation have a negative interaction effect on NEP in the Beijing-Tianjin-Hebei region. This study enhances the understanding of vegetation carbon sink dynamics in the region and provides a scientific basis for ecological management and climate change adaptation strategies.