Introduction Recently, immune cells within the tumor microenvironment (TME) have become crucial in regulating cancer progression and treatment responses. The dynamic interactions between tumors and immune cells are emerging as a promising strategy to activate the host's immune system against various cancers. The development and progression of hepatocellular carcinoma (HCC) involve complex biological processes, with the role of the TME and tumor phenotypes still not fully understood. Therefore, it is essential to investigate the importance of immune cell homeostasis in HCC. Additionally, understanding the molecular mechanisms and biological functions underlying tumor-immune cell interactions is increasingly recognized as vital for improving therapeutic outcomes in clinical settings. Methods A total of 790 HCC samples were selected from public databases and real-world independent clinical cohorts. Machine learning methods, focusing on immune-related indicators, were applied to these samples. The Boruta algorithm was employed to develop an ICI score, which was used to assess patient prognosis and predict responses to immunotherapy. Additionally, a new immune subtype analysis of HCC was performed. Cellular-level experiments confirmed the interaction between TME-related factors and the tumor microenvironment in HCC. To further validate the predictive power of the ICI score, a clinical cohort study was conducted at an independent clinical center. Results By evaluating immune gene expression levels, immune cell abundance, Immunescore, and Stromalscore, we initially identified three distinct immune subtypes of HCC, each showing significant differences in survival rates and heterogeneity. Subsequently, DEGs from 1022 immune subtypes were used to classify HCC samples into three immune genotypes, each characterized by distinct prognosis and tumor immune microenvironment (TIME) profiles. Furthermore, we developed the ICI score, a novel immunophenotyping method for HCC, which revealed significant variations based on gender, stage, progression, and DNA mutation profiles (p < 0.05). The ICI score also effectively predicted responses to immunotherapies, particularly through the chemokine signaling, focal adhesion, and JAK/STAT signaling pathways. Conclusion This research demonstrated that TME and immunophenotyping clusters can enhance prognostic accuracy for HCC patients. The independent prognostic indicators identified underscore the connection between tumor phenotype and the immune environment in HCC.