The Xinzhao gas zone of the Dongsheng gas field resides at the junction of three first-order tectonic units:Yishan slope,Yimeng uplift,and Tianhuan depression.Due to the influence of paleogeomorphology and provenance,different channel sediments vary significantly in this zone.The main target layer is the first member of the Shihezi Formation,which hosts a braided river sedimentary system,with tight reservoirs characterized by typical low porosities and permeabilities.The effective sandstone reservoir exhibits a small thickness,pronounced heterogeneity in gas content,and low-resolution seismic data,complicating sweet spot prediction.Hence,this study delved into the critical seismic prediction techniques for sweet spots in braided-river tight sandstone reservoirs.Firstly,the stratigraphic framework was established based on the three-dimensional seismic relative isochronous surface,characterizing the spatial distributions of channel sand bodies at different stages using the relative-spatial-resolution stratal slicing technique.Then,the seismic data were transformed from the time domain to the frequency domain using the wavelet transform time-frequency analysis technique.Based on the analysis of instantaneous spectrum differences in seismic data of different channels,the qualitative gas content prediction was achieved using the low-frequency energy ratio and the ABV absorption attribute,effectively supporting well deployment.Finally,the facies-controlled geostatistical inversion in the depth domain combining logs and seismic data was conducted for quantitative reservoir prediction, guiding the accurate design and optimization of horizontal well trajectories.The application of these techniques has increased the probability of penetration of gas reservoirs in the study area by seven percentage points, demonstrating satisfactory application effects.