In this review, genetic changes known to occur in human and experimental animal hepatocarcinogenesis are evaluated comparatively, with the aim of identifying genes that could potentially be targets of new preventive and therapeutic strategies, albeit the fact that although a step-by-step analysis of the premalignant stages has been largely accomplished in experimental hepatocarcinogenesis, this goal is still elusive in the case of humans. Overexpression of several of the genes implicated in the MAPK signaling cascade and cell cycle control appears to be most likely responsible for initiated cells acquiring a proliferating phenotype that facilitates the accumulation of structural changes in additional genes, resulting in the generation of autonomously growing preneoplastic and neoplastic lesions. Several gene abnormalities seen in precancerous lesions of rodents also occur in human hepatocellular carcinomas, suggesting that at least some of them could be present also in human precancerous lesions. Furthermore, there are reports that epigenetic events, such as abnormal DNA methylation, may be critical in hepatocarcinogenesis. DNA hypomethylation is an early event, both in human and experimental hepatocarcinogenesis, and its role in the activation of various genes, has been postulated. In recent years, linkage analysis studies have led to the identification of susceptibility/resistance loci that influence the progression stage of hepatocarcinogenesis in mice and rats. The relevance of these findings, though, will depend on the identification of the genes, and on whether in humans there are genes ortholog with rodent's susceptibility/resistance genes. It is proposed that rodent hepatocarcinogenesis represents a promising model for the identification of genes implicated in the early stages of the process, and that many of these genes may represent key targets for the application of gene therapy in the prevention and treatment of liver cancer.