Abstract: In order to generate a high-quality Digital Bathymetric Model (DBM) for the Penghu Channel, the differences among the DBMs in different sub-regions are analyzed and a new DBM is constructed by using adaptive regional spatial weighted fusion method. The results indicate that among the DBMs used most widely at the present the SPTM30_PLUS, GEBCO_2021 and SRTM15_PLUS are good in quality and have the Root Mean Square Error (RMSE) ranging from 7.931 to 8.016 meters, whereas the ETOPO1, GEBCO_2014 and TOPO V19.1 are poor in quality and have the RMSE ranging from 11.439 to 20.105 meters. The Penghu Channel digital bathymetric model constructed in the present study has a RMSE of 6.970 meters and its accuracy is respectively 12%, 13% and 12% higher than that of the GEBCO_2021, SRTM30_PLUS and SRTM15_PLUS. By using the DBM constructed by the method proposed in this study high-precision water depth points are restored and the topographic details are retained. The results of the study can provide a reference for reconstructing and updating timely the large-scale seabed topography from multiple datasets and can be promoted and applied to other environments and regions.