Gastric cancer (GC) has been an increasingly serious problem in public health. However, there is still a lack of efficient approach to diagnosis and treatment in time, especially in the field of targeted therapy. Increasing evidences demonstrated that DNA methylation plays an essential role in tumorigenesis and progression of GC. Thus the present study aims to identify DNA methylation-based prognostic biomarkers in GC. Two methylation array datasets (GSE25869 and GSE30601) and RNA-seq based gene profiling dataset (TCGA-STAD) were employed for exploring candidate DNA methylation-based biomarkers. Univariate Cox regression analysis was used to select the most efficient prognostic genes in GC patients. Weighted gene correlation network analysis (WGCNA) was performed to screen the cluster of co-expressed genes. As a result, our data proved that NRP1 was a hypomethylated / upregulated gene in GC tissues, and PDGFRB was strongly co-expressed with it. Both of them were significantly associated with the overall survival of patients. More importantly, high expression levels of NRP1 and PDGFRB were associated with malignant phenotypes in GC patients, including Laurén histological diffuse type and higher histological grade. Patients carrying high expression level of NRP1 and PDGFRB had a nearly two-fold increased death risk than others. In summary, the hypomethylated gene, NRP1, and its co-expressed gene, PDGFRB, were significantly correlated with tumor malignant phenotypes, which might serve as potential prognostic biomarkers for GC patients.