Abstract: Seismic indicators of gas hydrate, such as the BSR (Bottom Simulating Reflector), bright reflector under BSR and chimney, have been found in the Qiongdongnan area by seismic survey. There is significant correlation between the development of chimney and the distribution of BSR. Chimney is the main pathway for natural gas to migrate vertically to the HSZ (Hydrate Stability Zone), which offers favorable conditions for massive gas accumulation. Therefore, it is very important to reveal the development characteristics of gas chimneys in order to study of the accumulation and distribution of gas hydrate in the area. Differ from the traditional method to recognize gas chimney with seismic data based on columnar shape, weak reflection and other related attributes, this paper comprehensively studied the development characteristics of gas hydrate by neural network analysis technology and the correlation between gas chimney and the BSR.