In order to determine the nature and distribution of marine gas hydrate, a series of geophysical data processing techniques are used in this paper. By using the traditional seismic data processing, purpose seismic data-processing, wave impedance inversion techniques and geophysical well logging data processing based on Self-organizing feature map neural network, a great deal of useful information are abstracted to determine the gas hydrate zone beneath the seabed. The results show (1) the conventional seismic data processing techniques contains much useful information and provide a visual and qualitative evidence of the presence of gas hydrate in sediments under normal condition;(2)special processing techniques, such as attribute extraction and wave impedance inversion, is necessary so as to mine more effective data, they could compensate the shortage of conventional seismic data processing techniques used for distinguishing gas-bearing reservoirs;(3)as a kind of intelligent information processing technology, SOFM neural network is feasible for lithologic identification by logging data and has a high rate of identification of gas hydrate. In the end, the author hopes it may provide some useful clues to the exploration of gas hydrate.