The traditional gas detection method cannot meet the requirements of power department monitoring of oil-immersed electrical equipment at low cost and over a large range. Thus this paper presents a monitoring gas method. It uses the measured sound speed at three frequency points to get sound speed dispersion relaxation characteristics. First, the sound speed dispersion expression is simplified by deducing the theoretical formula of effective compressibility and the acoustic relaxation process. From the measured sound speeds at different frequency points, the sound speed dispersion is reconstructed, and the sound speed dispersion relaxation speed and frequency of the fault gas of oil-immersed electrical equipment are extracted. Finally, the gas composition is identified qualitatively and quantitatively using relaxation characteristics. Simulation results show that the reconstructed sound speed dispersion is consistent with the experimental results, and gas mixtures including CH4 and CO2 are identified. It provides a low-cost acoustic method with simple equipment for online detection of dissolved gas composition from oil-immersed electrical equipment. This work is supported by the National Natural Science Foundation of China (No. 62071189 and No. 62271451).