Abstract. The aim of the present study was to explore the potential mechanisms involved in each subtype of head and neck squamous cell carcinoma (HNSCC) via subpath analysis and to investigate their relevance in the prevention of HNSCC. Gene expression profiles of GSE6631 and GSE39366 containing 44 and 168 HNSCC samples, respectively, were downloaded from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) from samples in GSE6631 and GSE393666 were screened using the Detection of Imbalanced Differential Signal (DIDS) method respectively. DEGs in GSE39366 were matched with the DEGs in GSE6631 and were used to classify the subtypes of HNSCC based on hierarchical clustering analysis. Furthermore, DEGs were separated into different subtypes and then the pathway information was analyzed. The regulated miRNAs for the DEGs in each subtype were analyzed to select the significant subpaths. Totally, 1,095 DEGs from GSE6631 and 2,528 DEGs from GSE39366 were screened. Samples in GSE39366 were separated into four subtypes. Specific genes in each subtype and DEGs in the common gene set involved in a variety of pathways were identified. In addition, the significant miRNAtarget-pathway subpath of each subtype of HNSCC and the common gene set of HNSCC were also enriched. Our data suggest that human papillomavirus (HPV) is positively correlated with HNSCC in subtype 2. Several miRNAs (miRLet-7A, miR-1, miR-206, miR-153, miR-519A and miR-506) and their target genes (CYP46A1, BPNT1, MCM7 and COL5A1) are crucial for HNSCC prevention via different pathways and may provide further knowledge of the mechanisms involved in the progression of HNSCC.