Abstract: To evaluate the applicability of the adjustive full waveform inversion (AdFWI) to legacy shallow-water, narrow-azimuth, and short-streamer data, a velocity-model-building test for the X gas field in the central depression of the Yinggehai Basin was carried out. Because the legacy seismic data suffer from weak low-frequency energy, limited offsets, and narrow azimuthal coverage, the accuracy of velocity data for diapir-gas chimney system in the area is insufficient and the imaging is poor. Therefore, we conducted a velocity-update experiment. The data were acquired in water depths of 60～80 m and streamer lengths of 3～4 km. To make full use of local traveltime differences and reduce the risk of cycle skipping, adaptive full-waveform inversion was introduced based on diving wave tomography and common imaging point tomography velocity updates, starting from the lowest usable dominant frequency of 7 Hz and running for 10 iterations. Results show that, compared with the DWT and CIP tomography velocity models, the AFWI-updated velocity model exhibited a continuous low velocity response at the shallow gas chimney. In the corresponding prestack depth migration section, the continuity of reflection events in the blurred zone was enhanced, and the residual moveout in CIP gathers was reduced with improved gather flattening. This study provided a reference for gas chimney velocity model building, imaging evaluation, and reprocessing parameter selection in the central diapir belt of the Yinggehai Basin and the areas with similar shallow-water legacy data, and for delineating complex gas-cloud structural configurations, enhancing the reprocessing value of legacy data, and supporting natural gas exploration and development in the basin.