Objective To explore the inhibitory effects and mechanisms of gambogic acid (GA) on prostate cancer (PCa) using network pharmacology and model of heterogeneous transformation of prostate cancer. Methods. The potential target genes of GA in inhibiting the growth of PCa were obtained from PubChem, SwissTargetPrediction, SuperPred, SEA, GeneCards, OMIM databases and Venny 2.1.0. A protein-protein interaction network was constructed through the STRING database and Cytoscape 3.8.2. GO (gene ontology) and KEGG enrichment analyses were performed using DAVID and visualized to predict the targets and pathways of GA acting on PCa. A patient-derived tumor organoids (PDO) model of prostate cancer was constructed and combined with PCa cells (22Rv1, PC3, and DU145) to verify the results predicted by network pharmacology at the cellular level. CellTiter-Glo assay and CCK-8 experiment were carried out to observe the effects of GA on the viability of PDO and PCa cells, respectively. The changes in the levels of target proteins in cell lines and PDO after GA treatment were analyzed by qPCR and Western blot. Furthermore, a patient-derived xenografts (PDX) model of prostate cancer was constructed. After GA treatment, the tumor volume and weight were measured, and the expression changes of the targets predicted by network pharmacology in tumor tissues were analyzed by immunohistochemistry.Results Network pharmacology screening showed that the core target of GA in inhibiting the proliferation of PCa was STAT3, which was related to the HIF-1α signaling pathway. A theoretical framework related to the GA-STAT3-HIF-1α signaling pathway was preliminarily constructed. Further cytological experiments showed that the viability of 22Rv1, PC3, DU145 cells and PDO decreased after 48 hours of GA treatment, and the protein levels of HIF-1α, STAT3, and P-STAT3 were significantly down-regulated. In vivo experiments showed that the tumor volume and weight in the GA group were significantly reduced. Immunohistochemistry results showed that the expressions of STAT3 and HIF-1α in tumor tissues decreased after GA treatment. Conclusion Through the application of more clinically representative PDO and PDX models, combined with experimental studies on cell lines, the prediction results of network pharmacology were verified. GA showed a significant killing effect on PCa, and its mechanism of action may be related to the STAT3/HIF-1α signaling pathway.