OBJECTIVE: To systematically characterize tumor-associated macrophage (TAM) subsets in lung adenocarcinoma (LUAD) and establish a TAM-based prognostic risk signature for LUAD patients. METHODS: Single-cell RNA sequencing (scRNA-seq) and bulk transcriptomic data were integrated to identify TAM subsets linked to LUAD prognosis. Prognostic genes were screened using univariate Cox regression, refined via Least Absolute Shrinkage and Selection Operator (LASSO) regression, and used to construct a 10-gene risk signature. The signature's performance was validated in independent cohorts through receiver operating characteristic curves, Kaplan-Meier survival analysis, and a nomogram. Its predictive ability for immune checkpoint inhibitor (ICI) therapy response was assessed in the IMvigor210 and GSE78220 datasets. RESULTS: Six distinct TAM subpopulations were identified, with two subsets significantly correlated with poor prognosis. The 10-gene risk signature, derived from TAM-related genes, demonstrated strong prognostic performance in both training and validation cohorts. High-risk patients exhibited markedly worse overall survival compared to low-risk patients. Additionally, the signature effectively stratified patients based on their response to anti-PD-L1 therapy, with high-risk patients exhibiting reduced clinical benefit. A nomogram combining the risk signature with clinicopathological parameters further enhanced survival prediction accuracy, supporting its clinical applicability. CONCLUSION: This study established a novel TAM-based prognostic risk signature with robust predictive power for both survival outcomes and immunotherapy response in LUAD. These findings enhance our understanding of TAMs' clinical significance and provide a foundation for personalized immunotherapy strategies in LUAD.