Background: The intrinsic molecular subtypes of lung adenocarcinoma (LUAD) impact clinical treatment decision-making, but the molecular mechanisms are still unclear. Therefore, we aimed to identify sensitive biomarkers to evaluate LUAD patient prognosis. Methods: Differentially expressed RNAs from LUAD patients were obtained from The Cancer Genome Atlas (TCGA) database and they were used to construct a competitive endogenous RNA (ceRNA) network. Based on the examination of clinical data, long noncoding RNAs (lncRNAs) and mRNAs in the network were selected by univariate and multivariate Cox regression analysis. Finally, functional enrichment analysis was used to reveal prognostic signatures based on the classification into high and low-risk groups, survival analysis, and an independence test. Results: The ceRNA network consisted of 21 mRNAs, 53 lncRNAs, and 8 miRNAs that were selected from the differentially expressed RNAs identified. Next, based on univariate and multivariate Cox regression analysis, a prognostic signature, including two mRNAs (HOXA10 and CBX2) and four lncRNAs (LINC00460, LINC00330, DGCR5, and C14orf132) was constructed. Eventually, survival analysis showed that significant differences in survival rates between high and low-risk groups and the area under the curve (AUC) for three‐year survival was 0.714. Compared with clinical risk factors, including age, pathological stage, and TNM stage, our risk score had a higher prognostic value. Conclusion: By screening from a ceRNA network, we constructed a signature, including two mRNAs (HOXA10 and CBX2) and four lncRNAs (LINC00460, LINC00330, DGCR5, and C14orf132), that can be utilized as a prognostic biomarker in LUAD. This signature may provide options for clinical treatment.