The occurrence of hepatocellular carcinoma (HCC) related to liver cirrhosis is mostly accompanied by extensive immune infiltration. To reveal the infiltration immune cells landscape, single-cell RNA sequencing data from the healthy donor (HD), patients with liver cirrhosis (LC) and HCC were collected for analysis. By drawing a cell map and calculating the proportion of each cell type, total B cells were identified with a significant higher proportion in HCC (24.26%) than in LC (5.41%) and HD (5.82%), in which plasma cells account for 97.1% in HCC. While in HCC, TCGA datasets were taken for further investigation, and it was found that patients with lower proportion of plasma cells showed better prognosis. The pseudotime cell trajectory analysis of B cell population found that humoral immunity continuously changes during HD, LC and HCC, and humoral immune-related genes are highly expressed in the HCC stage. This suggests humoral immunity may play a key role in the development of LC-associated HCC. At the same time, single cell data of hepatocytes identified differentially expressed genes in HD/LC and LC/HCC groups, and a prognostic model constructed with six of the differential genes (FTCD, MARCKSL1, CXCL3, RGS5, KNG1, and S100A16) could classify HCC patients to two distinct risk groups (median survival time 2.46 years vs. 6.73 years, p < 0.001). Our study demonstrated the power of single-cell data analysis in dissecting tissues into infiltration and main body cells, it revealed the pivotal roles of humoral immunity infiltration in the landscape of HCC associated with cirrhosis, and the key tumor prognostic genes in hepatocytes themselves. These brought novel insights into studying microenvironment and tumor cells parallelly in cancer research. The interaction of both, rather than factors from one side may have caused tumorigenesis and progression.