Abstract: In the traditional seismic study, such as seismic data analysis, processing and attribute analysis, the Fourier transform based on stationary signal processing theory is the core technology. The traditional Fourier transform is a kind of integral transform, and seismic signals are usually non-linear and non-stationary, so it is difficult to reflect the partial feature of the non-stationary signal. However, the time-frequency analysis technique has the capability to reflect the partial feature of signal in time domain and frequency domain. This paper, based on the introduction to several common seismic time-frequency analysis methods, is devoted to in-depth study of adaptive time-frequency analysis method. Theoretical analysis and testing show that:the adaptive optimum Kernel time-frequency representation can repress cross-terms and has better time-frequency resolution, so it is worthy to do more research on seismic attributes extraction and spectral decomposition.