Non-stationer signal from the seismic survey is difficult to be directly interpreted in time domain analysis. Spectral decomposition is one of the spectral analysis methods that can analyze the non-stationer signal in frequency domain. The Fast Fourier Transform method was commonly used for spectral decomposition analysis, however, this method had a limitation in the scaled window analysis and produced pure quality for low-frequency shadow. The S-Transform and Empirical the Mode Decomposition (EMD) is another method of spectral decomposition that can be used to enhanced low-frequency shadows. In this research, comparison of the S-Transform and the EMD methods that can show the difference imaging result of low-frequency shadows zone is applied to Eldo Field, Jambi Province. The spectral decomposition result based on the EMD method produced better imaging of low-frequency shadows zone in tuning thickness compared to S-Transform methods.