Background: Hepatocellular carcinoma (HCC) has limited treatment options. Molecular analysis of its mutational landscape may enable the identification of novel therapies. Noninvasive strategies are critical to HCC given the challenge of obtaining liver tissue biopsies. Recent studies show that circulating tumor DNA (ctDNA) can reflect tumor genomic profiles. Thus, we aimed to analyze the molecular characteristics of HCC using ctDNA. Methods: A total of 308 plasma samples and 32 matched tumor tissues from Chinese patients with hepatocellular carcinoma (n = 308) were analyzed by Acornmed panel with 808 cancer-related genes. Results: A total of 4984 somatic variations were identified in 96.1% (296/308) ctDNA, including 4202 single nucleotide variants, 772 copy number variations and 10 fusions, and the median blood tumor mutation burden (bTMB) was 4.36 mut/Mb. Alterations in TP53 (152/308, 49.4%) were the most prominent variation, followed by those in TERT (48/308, 15.6%), ARID1A (37/308, 12.0%), CTNNB1 (34/308, 11.0%), NF1 (16/308, 7.7%), BRAF (16/308, 7.7%), PTEN (22/308, 7.1%) and EGFR (19/308, 6.2%). Interestingly, in MET and MYC, amplifications occurred more frequently than mutations. Moreover, we systematically evaluated clinically actionable mutations and totals of 587 potential drug-related targets from 211 patients (68.5%) were identified. In addition, in 32 patients with matched tumor tissues and plasma samples, sequencing detected alterations in 96.9% (31/32) tumor tissue DNA (tDNA) and 100% (32/32) ctDNA, respectively. Further analysis found that 25.4% of mutations of tissues were founded in ctDNA and 32.5% of mutations of ctDNA were detected in tissues. Among them, 68.8% (22/32) patients had at least one overlapping variant called by tDNA and ctDNA. The most common shared mutation was TP53 (8/32, 25.0%), followed by PTEN (2/32, 6.3%) and ARID1A (2/32, 6.3%). Meanwhile, two shared copy number variants were detected in two patients (6.3%), respectively. Conclusions: This study contributes to the understanding of the molecular characteristics of HCC using ctDNA and suggests potential clinical utility for this technology. Furthermore, it can more comprehensively reflect the molecular mutation status of patients by combining tissue and ctDNA detection.