The Sulige Gas Field is a typical fluvial tight sandstone gas reservoir. Its effective reservoirs are small in scale, and their superposition patterns and internal architectures are complex. Conventional single-scale multiple-point geostatistical methods have difficulty in reproducing large-scale spatial structures and local architectural details simultaneously. In this study, the Su-S Block of the Sulige Gas Field was selected as the study area. A multiresolution multiple-point geostatistical modeling method was developed for tight sandstone reservoir architecture. First, a geological knowledge base for braided-river reservoirs was established based on architectural-element identification, superposition-relationship analysis, and geometric-parameter statistics. Three-dimensional training images were then constructed to represent the spatial relationships among channels, braid bars, and internal interlayers. The multiresolution strategy was introduced into the Direct Sampling method, and reservoir architectures were simulated progressively from coarse to fine resolutions. A stratified modeling strategy was also adopted to characterize architectural differences among individual sublayers. The results show that the models generated by the proposed method are consistent with the geological understanding of the study area. Compared with conventional DS and Snesim, the MR-DS method provides better reproduction of channel distribution, braid-bar morphology, sandbody connectivity, and local architectural boundaries. The variogram results show that spatial structures in the training image are better preserved. The multidimensional scaling results indicate higher structural similarity and better simulation stability. The simulation time is also significantly reduced. The proposed method therefore provides an effective approach for three-dimensional architectural modeling of tight sandstone reservoirs and provides a reliable geological basis for refined characterization and development of remaining gas resources in the Sulige Gas Field.