Background: The basement membranes (BMs) are involved in tumor progression, while few comprehensive analyses to date are performed on the role of BM-related gene signatures in lung adenocarcinoma (LUAD). Thus, we aimed to develop a novel prognostic model in LUAD based on BMs-related gene profiling. Methods: The LUAD BMs-related gene profiling and corresponding clinicopathological data were obtained from the basement membrane BASE, The Cancer Genome Atlas (TCGA) and gene expression omnibus (GEO) databases. The Cox regression and least absolute shrinkage and selection operator (LASSO) methods were used to construct a BMs-based risk signature. The concordance index (C-index), receiver operating characteristic (ROC), and calibration curves were generated to evaluate the nomogram. The GSE72094 dataset was used to validate prediction of the signature. The differences in functional enrichment, immune infiltration, and drug sensitivity analyses were compared based on risk score. Results: In TCGA training cohort, 10 BMs-related genes were found, (e.g., ACAN, ADAMTS15, ADAMTS8, BCAN, etc). The signal signature based on these 10 genes was categorized into high- and low-risk groups regarding survival differences (p < 0.001). Multivariable analysis showed that the signature of combined 10 BMs-related genes was an independent prognostic predictor. Such a prognostic value of BMs-based signature in validation cohort of the GSE72094 were further verified. The GEO verification, C-index, and ROC curve showed that the nomogram had accurate prediction performance. The functional analysis suggested that BMs were mainly enriched in extracellular matrix-receptor (ECM-receptor) interaction. Moreover, the BMs-based model was correlated with immune checkpoint. Conclusion: This study identified BMs-based risk signature genes and demonstrated their ability to predict prognosis and guide personalized treatment of patients with LUAD.