Background: Numerous studies have used a multi-region sampling approach to characterize intra-tumor heterogeneity (ITH) in hepatocellular carcinoma (HCC). However, conventional multi-region sampling strategies cannot preserve spatial details of samples, and thus potential effects of spatial distribution of sampling sites on ITH estimation have long been overlooked. Methods: To address the problem of spatial information loss, a simple and easy-to-implement strategy named spatial localization sampling (SLS) has been proposed. Following SLS, we performed multi-region sampling on 14 HCC patients, collecting a total of 75 tumor samples with spatial information. Muti-omics sequencing was then performed on these samples. Results: Integrative analyses revealed significant correlation between spatial and molecular heterogeneity, implying that spatial sample distribution may influence ITH evaluation in HCC. Based on this discovery, a normalized diversity score was developed to quantify patient-wise ITH without the influence of sampling location bias, and it was then employed to categorize patients into low-ITH and high-ITH subclasses. The clinical and biological features of two ITH subclasses were systematically investigated, with the findings indicating that low-ITH tumors were associated with perivenous-type metabolic processes, activation of the Wnt signaling pathway and exclusion of T-cell infiltration. Also, we devised an approach for calculating gene-wise ITH, which contributed to the identification of a high degree of immune heterogeneity and the development of a low-ITH gene signature for risk prediction in HCC. Conclusions: Mounting evidence has shown that ITH is closely associated with survival and treatment outcomes in HCC patients. In this view, a better understanding of the properties of ITH is required to facilitate better clinical management of HCC. Here, we presented new approaches for measuring patient-wise as well as gene-wise ITH. Through these approaches, we revealed new biological features of low-ITH and high-ITH HCC tumors and demonstrated the existence of high immune heterogeneity in HCC. In addition, a prognostic signature was developed that had consistent prediction outcomes across different intra-tumor regions and could thus be used to predict the survival of HCC patients based on a single tumor biopsy sample.