Purpose The aim of the study is to identify a reliable gene panel to predict the prognosis of HNSCC patients by integrated genomic analysis. Methods Co-expression gene networks were constructed by WGCNA using GSE113282 gene expression profile. The biological functional investigation was performed by GO and KEGG function enrichment analysis. The hub gene module was screened by PPI. The prognostic gene panel was established by Lasso regression analysis, and further progression-free survival (PFS) analysis was validated by Kaplan-Meier survival analysis using GSE102995 data. Results We identified 195 genes associated with the overall survival (OS) status (correlation coefficients: - 0.42, and p value: 2e-05) by WGCNA. These genes were enriched in immune-related cytokines and pathways analyzed by GO and KEGG. Among the 195 genes, the module (42 genes) with the highest score was screened by PPI. A novel seven-gene predictive panel (CD19, CD40LG, CD5, CXCR6, FPR2, NCAM1, and SELL) was established by Lasso regression analysis, and the area under ROC curve (AUC) for 3-year OS status was 0.8298 and 0.7571, respectively, in the training set and the test set. The PFS time of the low-risk patients was significantly longer than the high-risk patients (p < 0.0001; log-rank test) by further validation using GSE102995 data. Conclusion The seven-gene panel may serve as a reliable predictive tool for HNSCC patients treated with platinum-based radio (chemo) therapy, and may be potential therapeutic targets for HNSCC patients.