Background: Acute myeloid leukemia (AML) is a highly aggressive hematological malignancy, and there has not been any significant improvement in therapy of AML over the past several decades. The mRNA vaccines have become a promising strategy against multiple cancers, however, its application on AML remains undefined. In this study, we aimed to identify novel antigens for developing mRNA vaccines against AML and explore the immune landscape of AML to select appropriate patients for vaccination. Methods: The genomic data and clinical information were retrieved from TCGA and GEO, respectively. The gene mutation data of AML were obtained from cBioPortal. GEPIA2 was used to analyze differentially expressed genes. The single cell RNA-seq database Tumor Immune Single-cell Hub (TISCH) was used to explore the association between the potential tumor antigens and the infiltrating immune cells in the bone marrow. Consensus clustering analysis was applied to identify distinct immune subtypes. The correlation between the abundance of antigen presenting cells and the expression level of antigens was evaluated using Spearman correlation analysis. The characteristics of the tumor immune microenvironment in each subtype were investigated based on single-sample gene set enrichment analysis. Weighted gene co-expression network analysis (WGCNA) was performed to identify co-expression modules and screen the hub genes. Results: Five potential tumor antigens were identified for mRNA vaccine from the pool of overexpressed and mutated genes, including CDH23, LRP1, MEFV, MYOF and SLC9A9, which were associated with infiltration of antigen-presenting immune cells (APCs). AML patients were stratified into two immune subtypes Cluster1 (C1) and Cluster2 (C2), which were characterized by distinct molecular and clinical features. Patients with C1 subtype showed higher expression level of the five candidate genes and worse prognosis compared with C2 subtype. Moreover, C1 subtype demonstrated an immune-hot and immunosuppressive phenotype, while the C1 subtype had an immune-cold phenotype. Furthermore, the two immune subtype showed remarkably different expression of immune checkpoints, immunogenic cell death modulators and human leukocyte antigens. Finally, WCGNA identified a module and five hub genes that can be used for predicting the response of AML patients to mRNA vaccine. Conclusion: CDH23, LRP1, MEFV, MYOF and SLC9A9 were potential antigens for developing AML mRNA vaccine, and AML patients in immune subtype 1 were suitable for vaccination.