DNA methylation (DNAm) plays an important role in the pathogenesis of psoriasis through regulating mRNA expressions. This study aimed to identify hub genes regulated by DNAm as biomarkers of psoriasis. Psoriatic skin tissues gene expression and methylation datasets were downloaded from Gene Expression Omnibus (GEO) database. Subsequently, multiple computational approaches, including immune infiltration analysis, enrichment analysis, protein–protein interaction (PPI) network establishment, and machine learning algorithm analysis (lasso, random forest, and SVM-RFE), were performed to analyze the regulatory networks, to recognize hub genes, and to clarify the pathogenesis of psoriasis. Finally, the hypermethylated genes were used to immune cell infiltration analysis, which revealed that psoriasis skin tissues were mainly composed of activated dendritic cells, resting mast cells, T follicular helper cells (cTfh), etc. Differentially expressed-methylated genes (DEMGs) were identified and partitioned into four subgroups and the 97 significantly hypermethylated and downregulated (hyper-down) genes accounted for the highest proportion (47%). Hyper-down genes were mainly enriched in glucose homeostasis, AMP-activated protein kinase (AMPK) signaling pathway, lipid storage disease, partial lipodystrophy, and insulin resistance. Furthermore, insulin receptor substrate 1 (IRS1), Rho guanine nucleotide exchange factor 10 (ARHGEF10) and retinoic acid induced 14 (RAI14) were identified as potential targets. These findings provided new ideas for future studies of psoriasis on the occurrence and the molecular mechanisms.