The reserve space of carbonate reservoir is complex and various in Puguang Gas Field, so the traditional hydrocarbon prediction methods can not get better results. The paper used the structural characteristics of seismic data volume to predict hydrocarbon reservoir, which took relationship between seismic data components on each seismic trace (that is structural characteristics of data volume) as study object to research the characteristics of composition and range of various parameters and their relationship with hydrocarbon bearing ability by detecting the amplitude value of each seismic trace, achieving the goal of predicting the hydrocarbon reservoir. Finally, the method was used to predict reservoir of Puguang Gas Field. The applied results demonstrated that the method can overcome the influence of reservoir nature and can get better results in the prediction of strong nonhomogeneous carbonate reservoir.