As the pollution of the atmospheric environment and the concern about health care of the public are increasing, the chemical gas sensor that can monitor the indoor air quality or the health status of the individual in real time is receiving great attention. In particular, chemiresistive gas sensors using metal oxides are the most promising type of gas sensors among the various types of gas sensor because they show high sensitivity and fast sensing ability even for ppm level gas molecules. However, it is still a major problem to selectively detect specific gases and accurately detect very low concentrations (ppb level) of gas molecules. In this Ph.D. thesis, I have developed a new synthesis method for synthesizing various metal oxide nanostructures and developed a platform for efficiently transferring nanoparticle catalysts to metal oxides in order to improve the sensing ability and selectivity of gas sensors. Furthermore, to solve the problem of selectivity which is a critical disadvantage of metal oxide, it is intended to improve sensitivity and selectivity of gas sensor based on metal oxide by introducing pore-tuned graphene oxide gas membrane as molecular sieving layer to selectively permeate the target gas molecules. By combining high surface area nanomaterials, nanocatalysts, and molecular sieving membrane, the accurate and high-resolution gas sensing system that can detect ppm level gas species was successfully achieved.