5-methylcytosine/N6-methyladenosine/N7-methylguanosine (m5C/m6A/m7G)-related genes play a critical role in tumor occurrence and progression, and non-apoptotic regulatory cell death (NARCD) is closely linked to tumor development and immunity. However, the role of m5C/m6A/m7G-related NARCD genes in hepatocellular carcinoma (HCC) remains unclear. We used m5C/m6A/m7G-related NARCD genes to construct a prognostic model of HCC for prognostic prediction and clinical treatment of patients. We obtained transcriptome data for HCC from The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC). Using the least absolute shrinkage and selection operator (LASSO) regression, we identified m5C/m6A/m7G-related NARCD genes and constructed a prognostic model through multivariate Cox regression. Model performance was assessed using Kaplan-Meier and receiver operating characteristic (ROC) curves, with external validation using the ICGC. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses were used to study differentially expressed genes between high- and low-risk groups. We also examined immune cell infiltration, drug response, and cell communication between tumor cells and immune cells in high-risk groups. We identified 140 m5C/m6A/m7G-related NARCD genes, using five of them to build the prognostic model. Functional enrichment analysis revealed enrichment in tumor and immune-related pathways for risk genes. The high-risk group displayed increased immune cell infiltration and better responses to immune checkpoint inhibitors (ICIs). High-risk patients were more responsive to cisplatin, doxorubicin, and mitomycin C, while low-risk patients were more sensitive to erlotinib. Cell communication analysis indicated that high-risk tumor cells used insulin-like growth factor (IGF) and macrophage migration inhibitory factor (MIF) signaling pathways to send signals to immune cells and received signals through the bone morphogenetic protein (BMP) and lymphotoxin-related inducible ligand (LIGHT) pathways. We have developed a prognostic model with m5C/m6A/m7G-related NARCD genes to predict the prognosis of HCC patients. This model can offer insights into the effectiveness of immunotherapy and chemotherapy for HCC patients.