Gas sensors play a crucial role in ensuring public safety, monitoring air quality, and detecting trace gases in industrial settings. The demand for gas sensors with high sensitivity, selectivity, efficiency, reliability, low cost, and low power consumption is substantial. Despite numerous proposals utilizing traditional metal oxide semiconductor materials, challenges persist in achieving satisfactory power consumption and selectivity. Herein, we address these challenges by developing smart gas sensors based on low-dimensional nanomaterials working at room temperature. Unlike their traditional counterparts, our sensors demonstrate improved selectivity and lower power efficiency. Combined with highly efficient machine learning algorithms, the low-dimensional nanomaterials-based gas sensors exhibit exceptional performance in both individual gas and complex odor identification. Our strategy paves the path to creating highly sensitive, selective, portable, and energy-efficient smart gas sensors using low-dimensional nanomaterials, catering to the growing needs of odor identification in various emerging fields.