Volatile organic compounds (VOCs) in exhaled human breath, key indicators of early-stage lung cancer, typically appear at ultra-trace concentrations in the parts-per-billion (ppb) range, making their detection under ambient conditions challenging. In this study, we employ first-principles density functional theory (DFT) combined with statistical analyses and the non-equilibrium Green’s function (NEGF) formalism to investigate the gas-sensing capabilities of MoS2, sulfur-vacancy-MoS2 (VS-MoS2), and oxygen-functionalized-sulfur-vacancy MoS2 (O2-VS-MoS2) toward three key VOCs, including 3-hydroxybutanone (3HB), isoprene (ISO), and ethylbenzene (ETB). We show that MoS2, VS-MoS2, and O2-VS-MoS2 all exhibit selectivity toward 3HB detection in the presence of major ambient gases (i.e., N2, O2, H2O, and CO2). Among them, O2-VS-MoS2 demonstrates the highest sensing performance, originating from the formation of robust interfacial H-O bonds arising from the orbital hybridization between the H-1s states of 3HB and the O-2pz orbitals of O2-VS-MoS2. This bond interaction induces pronounced modulation of the electronic structure, facilitating enhanced charge transfer and improved detection characteristics. Additionally, our quantum mechanical-derived microscopic findings enable the determination of macroscopic sensing metrics. The O2-VS-MoS2-based sensor achieves an ultrahigh 3HB detectability of 60.5 (216.5) ppb (ng/L) at room temperature, accompanied by an ultrafast response time of 1.24 × 10−1 ms. The current–voltage (I-V) characteristics further reveal a pronounced contrast upon gas exposure, confirming its exceptional sensitivity. These results position O2-VS-MoS2 as a highly promising platform for rapid, selective, and sensitive breath-based diagnostics.