<p>Cervical squamous cell carcinoma (CSCC) has an unfavorable prognosis with major therapeutic challenges. Natural killer (NK) cells play a pivotal function in anti-tumor immunity. However, the correlation between NK cells and heterogeneity and prognosis in CSCC lacks definitive understanding. This study seeks to elucidate the potential value of high-activity NK cell-related genes in prognosis and immunotherapy for CSCC. Transcriptome and single-cell sequencing data of people with CSCC were obtained from TCGA and EMBL-EBI databases, respectively. Single-cell data underwent quality control, dimensionality reduction, and identification of high-activity NK cells and their marker genes. After WGCNA application to screen NK-related genes. a prognostic risk model was constructed employing univariate Cox, LASSO Cox regression, and multivariate Cox regression analyses. The clinical implication of the model was validated through immune infiltration assessment, survival, gene set enrichment, tumor mutation analyses, and drug sensitivity prediction. High-activity NK cells and associated genes in CSCC were identified. A risk prognostic model based on high-activity NK-related genes was developed, yielding six key prognostic genes (RIPOR2, PTGER4, BIN2, MARCHF2, SPATA13, KLRC2). The model demonstrated robust predictive performance in training and validation sets. Patients in the low-risk group exhibited higher infiltration levels of NK, CD8<sup>&#43;</sup> T, and dendritic cells, along with increased sensitivity to immune checkpoint inhibitor therapy. Additionally, drug sensitivity analysis identified promising therapeutic candidates. This study, integrating single-cell and RNA sequencing, revealed the heterogeneity of NK cells in CSCC. The risk prognostic model provided prognostic biomarkers and therapeutic targets for CSCC patients, offering a theoretical foundation for immunotherapy research.</p>