Coalbed methane is highly enriched in southern Qinshui Basin’s P1s. In this work, we aim to predict Coalbed Methane in this region using frequency attenuation gradient. We first propose a time-frequency method to analyze wavelet transform, and then we build a geological model to simulate the real target reservoir. We successfully extract the time-frequency distribution of seismic signal. We demonstrate that using frequency attenuation gradient is possible to detect gas content in coal. Our results are consistent with the real logging results and better than reservoir prediction in terms of predication accuracy of coalbed methane. Our method can be used by domain scientist to predict coalbed methane in southern Qinshui Basin and other reservoir similar to Qinshui Basin.