Abstract: This study employed the α-glucosidase inhibitory activity model as an anti-diabetic assay and implemented a bioactivity-guided isolation strategy to identify novel natural compounds with potential therapeutic properties. Hypericum sampsonii was investigated, leading to the isolation of two highly modified seco-polycyclic polyprenylated acylphloroglucinols (PPAPs) ( 1 and 2 ), eight phenolic derivatives ( 3 – 10 ), and four terpene derivatives ( 11 – 14 ). The structures of compounds 1 and 2 , featuring an unprecedented octahydro-2H-chromen-2-one ring system, were fully characterized using extensive spectroscopic data and quantum chemistry calculations. Six compounds ( 1 , 5 – 7 , 9 , and 14 ) exhibited potential inhibitory effects against α-glucosidase, with IC₅₀ values ranging from 0.050 ± 0.0016 to 366.70 ± 11.08 μg·mL⁻¹. Notably, compound 5 (0.050 ± 0.0016 μg·mL⁻¹) was identified as the most potential α-glucosidase inhibitor, with an inhibitory effect about 6900 times stronger than the positive control, acarbose (IC₅₀ = 346.63 ± 15.65 μg·mL⁻¹). A docking study was conducted to predict molecular interactions between two compounds ( 1 and 5 ) and α-glucosidase, and the hypothetical biosynthetic pathways of the two unprecedented seco-PPAPs were proposed.