Abstract: Non-alcoholic steatohepatitis（NASH） has come to be a liver disease that seriously affects the quality of life and life expec-tancy of patients, and its pathogenesis is complex. To investigate the role of immune infiltration in the pathogenesis of NASH, data sets for hepatic simple steatosis（SS）, NASH, and healthy subjects were downloaded from the high-through put gene expression omnibus（GEO） database. Differentially expressed genes（DEGs） were identified by using R software limma package. The biological function was discussed by functional annotation and enrichment analysis. The immune infiltration types corresponding to clustering patterns were observed by using unsupervised consistency clustering and single sample gene set enrichment analysis（ssGSEA）. A total of 1 569 DEGs were obtained, and two modification patterns A and B were obtained via univariate regression evaluation and cluster analysis. The degree of hepatic fibrosis, lobular inflammation, ballooning degeneration and non-alcoholic fatty liver disease（NAFLD） activity score in pattern B were greater than these in pattern A. ssGSEA results showed that cells with immune response or pro-inflammatory function, as well as immunosuppressive cells were enriched in pattern B. Gene set variation analysis（GSVA） results showed that immune-related pathways such as cell adhesion molecules and metabolic pathways such as sphingolipid metabolism were enriched in pattern B. Finally, 5 core genes were determined by multiple stepwise Cox regression analysis to construct the hazard model. The AUC（area under curve） value of the training set and verification set of the prediction model were 0.982 and 0.987 respectively. These findings are helpful to have a deeper understanding of the pathophysiological mechanism of NASH and provide reference form edical prediction of the risk of NASH.