Lung adenocarcinoma (LUAD) is the most common type of non-small cell lung cancer and accounts for approximately 40% of all lung cancer cases. Multiple distant metastases are the major cause of mortality in lung cancer. In this study, single-cell sequencing datasets of LUAD were utilized to depict the transcriptome characteristic of LUAD based on the bioinformatic method. Firstly, the transcriptome landscape of heterogeneous cell types in LUAD was analyzed and memory T cells, NK cells, and helper T cells were revealed to be the common immune cells in tumor, normal, and metastasis tissue, respectively. Then, marker genes were calculated and 709 genes were identified to play a vital role in the microenvironment of LUAD. While macrophages were reported to act as one of the cells in LUAD, enrichment analysis of macrophage marker genes revealed the important role of macrophages in the activation of neutrophils. Next, the results of cell-cell communication analysis suggested that pericytes interact with broad immune cells via MDK-NCL pathways in metastasis samples, MIF-(CD74+CXCR4) and MIF-(CD74+CC44) interaction especially occurred between different cell types in tumor and normal samples. Finally, bulk RNA-seq was integrated to validate the prognosis effect of the marker gene and the maker gene of M2 macrophage, CCL20, showed the most related to LUAD prognosis. Besides, ZNF90 (Helper T cells), FKBP4 (memory T, helper T, Cytotoxic T, and B cells), CD79A (B cells), TPI1 (pericyte), and HOPX (epithelial cells, pericytes) were also pivotal in the pathology of LUAD, helping researchers understand the molecular insight of microenvironment in LUAD.