Time-frequency analysis is one of the important means for direct hydrocarbon detection with seismic data.Real seismic signals are non-stationary and shortcomings and deficiencies exist in conventional time-frequency analysis methods such as short-time Fourier transform

Gabor transform and wavelet transform.To tackle these problems

a new time-frequency analysis method-optimal generalized S-transform-is achieved by adding a new parameter "p" in S-transform to adjust the standard deviation of window function and selecting the optimal value of p according to the measurement criteria of the time-frequency aggregation proposed by Jones and Parks.Trial calculation with synthetic signals shows that this method has a better time-frequency aggregation than conventional time-frequency analysis methods.Application of this method to actual seismic data shows that the frequency-division data volume generated by this method is effective in hydrocarbon detection and identification.