Abstract: In the deepwater lacustrine fan systems of the lower second member of the Dongying Formation in the eastern Bonan Low Uplift, Bozhong Sag, Bohai Bay Basin, the absence of well control poses significant challenges to understanding genetic mechanisms and predicting reservoir sand-richness. Based on 3D seismic and well-logging data from peripheral wells, this study develops an integrated workflow that combines artificial intelligence with seismic sedimentology. First, a sand-richness index (FSI) is constructed from well logs to quantitatively characterize lithology and reservoir quality. Second, a dual-criterion strategy combining Pearson correlation analysis and Random Forest feature importance is employed to select three seismic attributes most sensitive to FSI: sweetness, neighboring trace coherence, and total energy. Third, a Random Forest regression model is established to quantitatively map the relationship between these attributes and FSI. Finally, a virtual well (D) is deployed in the deep sag area to predict the vertical lithological succession of the fan, and seismic sedimentology techniques are applied to delineate its planar distribution and explore controls on sand/mud accumulation. Results indicate that: (1) FSI effectively discriminates sandstone from mudstone with an accuracy exceeding 87%, and the established model demonstrates high predictive performance (R²=0.84); (2) the lacustrine fan is characterized by a "mud-enveloped-sand" architecture, with a sand-to-mud ratio below 20% and a typical vertical succession of thick mudstone - thin sandstone interbedded with thick mudstone - thick mudstone; (3) the mud-rich nature of the fan is preliminarily interpreted to result from the coupled effects of structural slope-break zones, high mud-prone provenance supply, and density-flow differentiation. The proposed workflow, integrating geological mechanism constraint, AI-driven analysis, and seismic sedimentological interpretation, provides a feasible approach for reservoir prediction and exploration risk mitigation in deepwater, well-constrained settings of continental rift basins.