Abstract: To explore the possible molecular mechanism of Shugan Jiangzhi granules in regulating ferroptosis in the treatment of non-alcoholic fatty liver disease（NAFLD） by using the method of bioinformatics. The potential intersection targets of "Shugan Jiangzhi granules-ferroptosis" were obtained by network pharmacology. The differentially expressed genes（DEGs） of the intersection target of "Shugan Jiangzhi granules-NAFLD-ferroptosis" were screened based on the GSE89632 dataset, and the correlation analysis, enrichment analysis and immune infiltration analysis were performed by using R4.2.2 software. The samples of GSE89632 dataset were clustered and typed, the machine learning model was constructed to screen the key genes of DEGs, and the risk prediction nomogram model was constructed and verified. The clinical correlation of key genes was analyzed. Finally, a total of 18 DEGs of "Shugan Jiangzhi granules-NAFLD-ferroptosis" were obtained. DEGs enriched in Th17 cell differentiation, thyroid hormone signaling pathway, ErbB signaling pathway, IL-17 signaling pathway and so on. Random forest（RF） model is the optimal machine learning model. AURKA, SREBF1, HMOX1, MYC and JUN were the top five genes in the RF model. HMOX1 and JUN were positively correlated with body mass index（BMI）. Nomogram model has good predictive ability for the risk of NAFLD. This study suggests that Shugan Jiangzhi granules can regulate ferroptosis to treat NAFLD through network regulation of AURKA, SREBF1, HMOX1, MYC, JUN and other differentially expressed ferroptosis-related genes, and mediated immune regulation, Th17 cell differentiation, thyroid hormone signaling pathway, ErbB signaling pathway, IL-17 signaling pathway. The nomogram model based on AURKA, SREBF1, HMOX1, MYC and JUN can not only accurately diagnose NAFLD, but also indirectly evaluate the efficacy of Shugan Jiangzhi granule in the treatment of NAFLD.