Graphene-based nanomaterial was widely used in gas sensors, detection, and separation. However, weak adsorption and low selectivity of the pristine graphene used for gas sensors are major problems. Here, using density functional theory (DFT) calculations, we reported the significant increase of four gas molecules (N2, CO2, C2H2, and C2H4) adsorption on the transition metal ion (Fe3+, Co2+, Ni2+)-modified graphene complex (Fe3+/Co2+/Ni2+-G) comparing to be absorbed on the pristine graphene (G). Moreover, the Co2+-G is suitable for the selective separation of C2H4/C2H2 due to the larger adsorption energy difference (8.5kcal/mol) between them. The addition of transition metal ions also decreased the HOMO-LUMO gap of the systems, which benefits the enhancement of electrical conductivity. This suggests that the transition metal ion-modified graphene can be used to distinguish the different gas molecule's adsorption, facilitating the design of graphene-based gas sensors and selective separation. All the density functional theory (DFT) calculations were performed by B3LYP with the GD3 dispersion method using Gaussian 16 software. The basis set 6-31G(d) was used for C, H, O, and N atoms, and Lanl2DZ was used for transition metal ions (Fe3+, Co2+, Ni2+). The DOS analysis and energy decomposition analysis were performed using the Multiwfn program.