〔Abstract〕 Objective To screen for potential biomarkers of anal fistula(AF) by using weighted gene co-expression network analysis(WGCNA), machine learning, immune infiltration analysis, and animal experiments. Methods Download transcriptome data from the gene expression omnibus containing AF and peri-fistula tissue(PF) for differential analysis. Gene ontology(GO) and Kyoto encyclopedia of genes and genomes(KEGG) pathway enrichment analyses on differentially expressed genes(DEGs) were performed. WGCNA results were integrated with DEGs to screen for genes related to AF. Machine learning methods such as the least absolute shrinkage and selection operator(LASSO), support vector machine recursive feature elimination(SVM-RFE), and random forest(RF) were utilized to screen potential biomarkers for AF. Immune infiltration analysis was conducted and the AF rat model was replicated for validation. Results A total of 377 DEGs were obtained, mainly enriched in pathways such as B cell receptor signaling and chemokine signaling. Machine learning algorithms identified matrix metalloproteinase 13(MMP13) as a potential biomarker for AF. In AF samples, memory B cells, plasma cells, M0 macrophages, and M1 macrophages were higher than in PF samples, while resting CD4 memory T cells and resting dendritic cells were lower than in PF samples. MMP13 showed a positive correlation with M0 macrophages, activated mast cells, and immature B cells, and a negative correlation with resting mast cells. Experimental results showed that MMP13 expression levels were higher in rat AF samples compared to the control group. Conclusion The onset of AF involves various immune cells and signaling pathways. MMP13 is significantly upregulated in AF tissue and correlates with multiple immune cells, making it a potential novel biomarker of AF.