A new tri-ruthenium dihydrido cluster, [Ru3(μ-H)2(CO)6(HDCQX)2] (1), in which two HDCQX (H2DCQX = 6,7-dichloroquinoxaline-2,3-dione) ligands cap both faces of the tri-ruthenium triangle similarly through oxygen atoms, has been synthesized and characterized by mass, IR, NMR (1H and 13C) and electronic spectroscopy. DFT calculations, in both gas and solution (CPCM/DMSO) phases, were performed to analyze the electronic structure of the cluster and interpret its vibrational and NMR spectra. The presence of non-equivalent bridging hydrides was assigned by 1H NMR as two singlet resonances at δ −16.85 and −17.67 ppm and confirmed by theoretical DFT-GIAO calculations at δ −12.92 and −14.48 ppm. Time-dependent density functional theory (TD-DFT) calculations, performed to characterize the frontier orbitals and interpret the electronic spectrum, indicated that the lowest-energy transition of 1 originates from a highly delocalized HOMO with Ru–Ru (σ), Ru–CO (π) and Ru–CO (σ*) contributions and ends in the LUMO with Ru1–Ru2–Ru3 (σ*) and QX/QX′ (π*) anti-bonding character. The natural transition orbital (NTO) corresponding to this transition revealed a decrease of the electron density on the Ru1Ru2(CO)4 fragment and an increase of the electron density on the Ru1–Ru2–Ru3 unit and the HDCQX ligands. The Ru–Ru and Ru–ligand bonding was studied by the quantum theory of atoms in molecules (QTAIM) and electron localization function (ELF) methods. Both confirmed the presence of unique bonding between the Ru1 and Ru2 atoms and not between Ru3 and the other Ru atoms. While the QTAIM assigned the Ru1–Ru2 bonding as a transit between pure closed-shell and pure shared-shell interactions, the ELF analysis suggested that this bond is dominated by a fluctuation of electron density. The covariance analysis indicated that the core basin populations C(Ru1) and C(Ru2) account for 76% of the V(Ru1, Ru2) variance. Natural bond orbital (NBO) analysis of the Ru3(µ-H)2O4(CO)6 unit indicated a 2c–2e bonding type for the Ru–O, Ru–C and Ru–Ru bonds with orbital occupations of 1.88–1.95, 1.95–1.96, and 1.57 e, respectively, and a 3c–2e bonding for the Ru–H–Ru bonds with an orbital occupation of ca. 1.7 e. NBO second-order perturbative energy analysis was performed to estimate the strength of the Ru–CO π-back-donation and to establish a correlation between the stabilization energy (E(2)) and the spectroscopic data. Although the calculated stabilization energy (E(2)total) follows the expected trend of decreasing the CO stretching frequency (νCO) and increasing the 13CO chemical shift with the extent of the Ru–CO π-back-donation, the latter parameter was more consistent in estimating the strength of such interactions. The in vitro anticancer activities of the title compound and cisplatin were determined against three common types of human cancer cell lines, HCT-116 (colon), HELA (cervix) and MCF-7 (breast). The IC50 values of 3.18 ± 0.40, 4.21 ± 0.29 and 5.39 ± 0.52 μM, respectively, represent 4- and 2-times improvement in potency relative to cisplatin (12.06 ± 1.28, 10.34 ± 2.08 and 6.29 ± 1.68 μM, respectively) against the HCT-116 and HeLa cancer cell lines.