Gas detection is attaining significant attention in the healthcare sector due to the health issues involved with the inhalation of toxic gases present in the air. Though the existing sensors can efficiently detect the target gases, they suffer from high power consumption, interference with other gases, and baseline drift. Here, we present a drift-independent technique to achieve selectivity and low power consumption with a single unfunctionalized metal oxide gas sensor. We utilized adsorption-desorption noise to discriminate H2S from NH3 and CO2. As it addresses typical challenges of existing resistive gas sensors, the proposed methodology paves the way for developing future low-power selective gas sensing technology. © 2021 Elsevier B.V.