Lung cancer is the most common malignancy worldwide. This study aimed to identify miRNA biomarkers of lung adenocarcinoma and to investigate their molecular mechanisms. miRNA expression profiling of tumor tissues and adjacent normal tissues from 10 patients were detected using microarray. Differentially expressed miRNAs (DEMs) were identified, and were verified using quantitative reverse transcription-PCR. Thereafter, correlations between DEM expression and clinicopathological features were determined in 49 patients. Furthermore, Targetscan was utilized to predict target genes, among which transcription factors (TFs) were identified. The interactions among miRNAs, TFs and target genes were used to construct an miRNA-TF-target network. Totally, 11 DEMs were identified, among which two downregulated miRNAs (miR-126-3p and miR-451a) were validated. Low levels of miR-126-3p and miR-451a were associated with poor pathological stage, large tumor diameter and lymph node metastasis (P<0.05). Receiver operating characteristic analysis showed that both miRNAs could predict pathological stage, tumor diameter and lymph node metastasis of lung adenocarcinoma (AUC >0.65, P<0.05). For miR-126-3p, 154 target genes were predicted (e.g., PLXNB2), which were enriched in 29 pathways mainly concerning apoptosis and cancer. For miR‑451a, 397 target genes were predicted, which were enriched in 5 pathways including 'PPAR signaling pathway'. Ten genes were co-regulated by miR-126-3p and miR-451a, e.g., TSC1. Furthermore, an miRNA-TF-target network was constructed, and a sub-network was identified, including 2 miRNAs, 15 targets, and 7 TFs. In conclusion, miR-126-3p and miR-451a predicted the severity of lung adenocarcinoma. However, the possible mechanisms explored by bioinformatics need to be further validated.