Human hepatocarcinogenesis is a complex process with many unresolved issues, including the cell of origin (differentiated and/or progenitor/stem cells) and the initial steps leading to tumor development. With the aim of providing new tools for studying hepatocellular carcinoma initiation and progression, we developed an innovative model based on primary human hepatocytes (PHHs) lentivirus-transduced with SV40, HRAS with or without hTERT. The differentiation status of these transduced-PHHs was characterized by RNA sequencing (including lncRNAs), and the expression of some differentiation markers confirmed by RT-qPCR and immunofluorescence. In addition, their transformation capacity was assessed by colony formation in soft agar and tumorigenicity evaluated in immune-deficient mice. The co-expression of SV40 and HRAS in PHHs, in association or not with hTERT, led to the emergence of transformed clones. These clones exhibited a poorly differentiated cell phenotype with expression of stemness and mesenchymal-epithelial transition markers and gave rise to cancer stem cell subpopulations. In vivo, they resulted in poorly differentiated hepatocellular carcinomas with a reactivation of endogenous hTERT. These experiments demonstrate for the first time that non-cycling human mature hepatocytes can be permissive to in vitro transformation. This cellular tool provides the first comprehensive in vitro model for identifying genetic/epigenetic changes driving human hepatocarcinogenesis.