Considering “carbon sources” and “carbon sinks”, this study is aimed at exploring the direct impacts and spatial spillover effects of urbanization (UR), ecological construction (EC) and their interaction on land use carbon emissions (CE) and land use carbon absorption (CA). This investigation contributes to resolving the dilemma posed by the need to balance economic development and ecological preservation in the context of energy conservation and emission reduction. Additionally, it is aimed at facilitating the achievement of dual carbon goals and promoting green and sustainable development. To accomplish this, the focus is placed on 279 Chinese cities during the period from 2005 to 2020. A comprehensive approach that combines multisource remote sensing imagery and socioeconomic data is employed to construct the core indicators. The spatial Durbin panel model is then utilized to analyze the direct effects, spatial spillover effects, and heterogeneity. The findings are as follows: (1) Both CE and CA in China exhibit significant spatial clustering. (2) UR has a significant promoting effect on CE, but it exhibits an inhibitory effect on CA and on CE in spatially correlated regions. (3) The ecological restoration intensity (RESI) and level of environmental regulation (ER) exacerbate the pollution haven effect, leading to a weakening of CE in the local region and the deterioration of the environment in its spatially correlated areas. Interestingly, RESI has a significant positive impact on local CA, while ER shows a substantial negative influence. (4) The interaction between UR and RESI weakens the CE in both the local and spatially correlated regions, resulting in a diffusion effect. However, the interaction between UR and ER triggers a green paradox effect. In the case of CA, these two interaction effects yield distinct outcomes in the local region and neighboring areas. It is worth noting that these effects are heterogeneous across city size, type, and location.