The early stages of hepatocarcinogenesis in human chronic liver diseases are characterized by the emergence of preneoplastic lesions of which some will eventually develop into hepatocellular carcinoma (HCC). Basic studies on the genetic and epigenetic alterations of these preneoplastic lesions may eventually lead to new therapeutic strategies. Clinicopathological studies are also important in order to determine optimal management of patients with a preneoplastic lesion. This article aims to provide a comprehensive review of the current concepts of preneoplastic lesion in chronic liver diseases. The microscopical small-cell dysplastic focus is the smallest morphologically recognizable precursor lesion of HCC and therefore is a logical target of study to elucidate the earliest events in hepatocarcinogenesis. In contrast, large-cell dysplasia is not a precursor lesion, but appears to be of clinical value because of its good predictive value for development of HCC. Dysplastic nodules (DNs) are macroscopically recognizable precursor lesions of HCC and high-grade DNs (HGDNs) have a risk of malignant transformation. Detection of DNs and correct differentiation from small HCC (<2 cm) is sometimes difficult, especially when only imaging techniques are used. Additional clinicopathological studies on identification and optimal treatment of DNs are necessary. Molecular studies on HGDNs and small HCCs may yield much information on the genetic mechanisms involved in the transition from severe dysplasia to early malignancy. In contrast, currently available data indicate that (large) regenerative nodules do not represent a distinct step in hepatocarcinogenesis. Animal models will be helpful in the further unravelling of human HCC development, provided that studies are performed on models that are good representatives of human hepatocarcinogenesis. We propose three criteria by which good mimickers can be identified.