Pancreatic adenocarcinoma (PAAD) is a malignant tumor of the digestive system, which develops rapidly and has no obvious early symptoms. This study aims to discover the biomarkers associated with PAAD development. We obtained RNA expression of PAAD patient samples and corresponding clinical data from The cancer genome atlas (TCGA), and screened out BMP/RA-inducible neural-specific protein 2 (BRINP2) gene which is highly associated with PAAD severity. Then, gene ontology (GO) enrichment, Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis and single-sample gene set enrichment analysis (ssGSEA) analysis were performed to explore the biological functions of BRINP2. Subsequently, long non-coding RNA (lncRNAs) associated with BRINP2 were screened out via correlation analysis, and Cox regression analysis and least absolute shrinkage selection operator (LASSO) regression analysis were used to construct the risk prediction model. We further validated the expression level of BRINP2 and its associated lncRNAs in BRINP2-associated lncRNAs prognostic model <i>in vitro</i>. We proposed that BRINP2 might be correlated to the tumor immune microenvironment and could also be used as a biomarker for PAAD progression. GO enrichment analysis and KEGG pathway analysis showed that the prognostic model was highly correlated to immune microenvironment-related pathways. Additionally, we established a BRINP2-associated lncRNAs prognostic model consisting of three lncRNAs. We validated the expression trends of BRINP2 and its associated lncRNAs in BRINP2-associated lncRNAs prognostic model in PAAD cells with various severity of metastatic potential using the quantitative real-time PCR (qRT-PCR). Meanwhile, <i>pRRophetic</i> R package was employed to predict potential therapeutic drugs for BRINP2-associated lncRNAs prognostic model of PAAD. The results suggest that BRINP2 can be used as a novel prognostic biomarker for PAAD.