Soil microorganisms are both involved in the decomposition of soil organic carbon and the driver of soil organic carbon transformation and fixation, affecting soil carbon cycle and stability. However, it remains unclear how conservational tillage affects soil respiration and carbon cycling regulated by soil C-cycling functional genes. In this study, based on the conservation tillage experiment trial of the Northeast Institute of Geography and Agricultural Ecology, Chinese Academy of Sciences, we focused on the effects of three tillage treatments including no-tillage (NT), moldboard plow (MP), and conventional tillage (CT) under the continuous maize system on the soil CO₂ efflux, Carbohydrate-Active enzymes (CAZy) related gene, and functional gene compositions related to carbon cycle (carbon fixation, methane metabolism, and carbohydrate metabolism) by shotgun metagenomic sequencing. Soil CO₂ efflux based on a 6-year average under NT was significantly lower than that under MP and CT in the early time of the growing season. Specifically, NT decreased soil CO₂ efflux over MP by 28% (May), 11% (June), 23% (July), and decreased by 31% (May), 19% (June), and 7% (July) over CT. CAZy database annotated results showed that tillage treatments significantly affected some CAZy families such as glycoside hydrolases (GH102, GH5_38, and GH13_17), glycosyl transferases (GT39), and polysaccharide lyases (PL17 and PL5_1) with the higher total abundances of these differential CAZy genes in NT and MP than that in CT. Moreover, there were divergences in soil C-cycling functional gene composition under different tillage regimes, and the gene compositions related to carbon fixation, methane metabolism, and carbohydrate metabolism under NT were different from that under MP and CT based on the KEGG database. The relative abundance of up-regulated carbon fixation functional genes (the average of the relative abundance of all up-regulated functional genes) in NT soil was higher than that of CT and MP with 17% and 11%, while the down-regulated genes abundance (K01007 and K00170) under NT were lower than these of CT and MP by 19%, 14% and 21%, 17%, respectively. The relative abundance of up-regulated methane metabolism genes in NT soil was higher than that of CT and MP by 15% and 10%, the down-regulated gene abundance was lower by 13% (CT) and 11% (MP), respectively. The relative abundance of up-regulated carbohydrate metabolism gene in NT soil was higher than that in CT and MP by 23% and 14%, the down-regulated gene abundance was lower by 25% (CT) and 18% (MP), respectively. Soil bulk density and dissolved organic carbon (DOC) were the driving factors of the functional gene composition of soil carbon cycle shown by db-RDA analysis, and the up-regulated genes related to carbon fixation, methane metabolism, and carbohydrate metabolism under NT were significantly positively related to soil DOC, bulk density or soil water content. Furthermore, the correlation analysis showed that soil CO₂ efflux was significantly correlated with the functional gene composition of the soil carbon cycle, especially with some soil carbohydrate metabolism functional genes. Our results suggested that NT affected the soil C-cycling functional genes by changing soil physicochemical properties, and the overlays of residues returning and reducing interference in NT could improve the soil carbon sequestration potential by shaping the soil C-cycling function gene composition.