The seeds of Sterculia lychnophora Hance, commonly known as Pangdahai (PDH) in Chinese, have found extensive use in both culinary and traditional medicinal practices. However, a comprehensive understanding of the chemical composition of PDH has been lacking. This study proposes a strategy that integrates biosynthetic pathway analysis with feature-based molecular networking (FBMN), aiming for a thorough and global characterization of the chemical compositions of PDH. The FBMN map reveals potential compounds with structural similarity, and the MS/MS fragments could be annotated based on library matches, which could predict the plausible biosynthetic pathways in PDH, accomplishing the annotation of compounds clustered in FBMN by integrating biosynthetic pathways. Consequently, 126 compounds were plausibly or unambiguously identified, including 37 phenolic acids and glycosides, 20 flavonoids and glycosides, 12 procyanidins, 21 alkaloids, 22 lipids, and 14 others. Leveraging the information, 40 compounds, including 1 unique isoquinoline alkaloid and 2 rare linear furocoumarins, were isolated and confirmed. This study not only demonstrates a highly effective approach for identifying compounds within complex herbal mixtures but also establishes a robust foundation for the further development of PDH.