Objective In this study, we aimed to identify the key microRNAs (miRNAs) and potential target genes through bioinformatics analysis, and investigate the underlying mechanisms of sepsis.Materials and Methods We collected miRNA expression profiles from sepsis patients and healthy individuals, screened differentially expressed miRNAs (DEMs) between sepsis patients and healthy individuals by bioinformatics analysis, and constructed miRNA-mRNA regulatory networks using online databases. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was used to annotate the biological functions and pathways of the genes. Single Sample Gene Set Enrichment Analysis (ssGSEA) assessed immunological characteristics in sepsis samples. Single cell sequencing (scRNA-seq) data were used to discover gene expression in different cell clusters.Results Four miRNAs were significantly differentially expressed in sepsis patients compared to healthy controls, with hsa-miR-34b-3p, hsa-miR-3663-3p and hsa-miR-4446-5p upregulated and hsa-miR-625-5p downregulated. ADRB2 may be a potential target of hsa-miR-34b-3p, and DisGeNET database showed that ADRB2 may be related to sepsis. Receiver operating characteristic (ROC) analysis suggested that ADRB2 has potential as a diagnostic marker for sepsis. The ssGSEA result showed that ADRB2 expression was positively correlated with T cell co-inhibition, and negatively correlated with dendritic cell infiltration. ScRNA-seq data showed that ADRB2 expression was increased in natural killer (NK) cells and natural killer T (NKT) cells in sepsis patients in contrast to healthy controls.Conclusion ADRB2 may suppress the autoimmunity of patients with sepsis, thus aggravating sepsis. It can be used as a new diagnostic biomarker and molecular therapeutic target. Hsa-miR-34b-3p can inhibit the expression of ADRB2, relieve its immunosuppressive effect and alleviate sepsis to a certain extent.