BACKGROUND: Head and neck squamous cell carcinoma (HNSCC) is a prevalent malignancy often diagnosed in advanced stages. Despite advancements in therapy, it retains a high mortality rate and significant recurrence risk. This study utilizes single-cell sequencing (scRNA-seq) to unravel HNSCC's complexity, identify therapeutic targets, and refine prognostic models. METHODS: Pseudotime trajectory and stemness analyses were performed on HNSCC tumor subpopulations, focusing on the C2 MALAT1+ Tumors subpopulation, which had the lowest CytoTRACE Score and represented the Lineage 2 endpoint in Slingshot analysis. The study examined programmed death and metabolic pathways in each subpopulation and developed a novel prognostic model using LASSO regression. RESULTS: The C2 MALAT1+ Tumors subpopulation exhibited reduced expression of programmed death pathways (e.g., Entotic cell death, Apoptosis, Pyroptosis) and metabolic pathways (e.g., Riboflavin metabolism, Glycolysis/Gluconeogenesis). Key transcription factors included LEF1, RFX3, CREM, MZF1, and ZNF202. Prognostic models based on the MALAT1 Tumors Risk Score (MTRS) revealed worse survival and higher tumor purity in the high MTRS group. Risk genes included ADM, RPL31, EIF5B, and TAF7. Additionally, activated CD4 memory T cells were enriched in the high MTRS group, which also showed greater sensitivity to Cisplatin, Docetaxel, and Paclitaxel. CONCLUSIONS: ScRNA-seq revealed the heterogeneity of HNSCC subpopulations, highlighting the unique features of the C2 MALAT1+ Tumors subpopulation. This study identified novel prognostic markers and therapeutic targets, offering insights into HNSCC progression, drug resistance, and potential treatments.