Accurate mapping of karst features is crucial for the resources exploration and recovery. The 3D seismic data can offer some insights in the spatial distribution of the karsts, but with low resolution, since the boundary of karst caves cannot be clearly delineated and the small-scale kart features are smeared in the seismic reflections. I propose to use time-frequency decomposition via a frequency interval analysis approach to accurately map the subsurface karst features with high resolution and accuracy. The time-frequency resolution can affect the final depicted result, and I suggest using the high-resolution synchrosqueezing wavelet transform (SSWT) to perform such task. The frequency interval analysis approach can provide us with more freedom in controlling the resolution and reliability of the final depicted result, compared with the traditional single frequency slice based approach. The proposed method is easy to implement and deserves more attention to assist in geological research on the topic of karst evolution.