Glioblastoma multiform (GBM) is the most common, most invasive, and malignant type of primary brain tumor with poorly prognosis and poorly survival rate. Using GSE22891 the expression and methylation status of same GBM patients was evaluated to explore key epigenetic genes in GBM. Using |log2FC| > 1 and FDR < 0.05 as the threshold, DEGs including 4910 downregulated and 2478 upregulated were screened and by |log2FC| > 0.2 and p.value < 0.05, 3223 DMCs were detected. By merging the results of DEGs and DMCs, 643 genes were selected for network analysis by WGCNA, and based on expression values three modules and by methylation values, one module was selected. Using STRING and Cytoscape, PPI network of genes of all modules were constructed separately. According to the PPI network, core genes were picked out. The expression status of core genes was evaluated using GSE77043, GSE42656, GSE30563, GSE22891, GSE15824, and GSE122498, and 50 genes were validated. The methylation status of 50 genes was explored using GSE50923, GSE22891, and GSE36245, and finally, 12 hub genes including ARHGEF7, RAB11FIP4, PPP1R16B, OLFM1, CLDN10, BCAT1, C1QB, C1QC, IFI16, NUP37, PARP9, and PCALF were selected. Using GEPIA database, the expression and survival plot, and using cBioportal the scatterplot of methylation versus expression of 12 hub genes were extracted based on TCGA. To determine the diagnostic values of the hub genes, the ROC curve and the area under the curve (AUC) were extracted based on GSE22891.