It is crucial to unravel molecular determinants of responses to immune checkpoint blockade (ICB) therapy, because only a small subset of advanced non-small cell lung cancer (NSCLC) patients responds to ICB therapy. Previous studies were concentrated on genomic and transcriptomic markers (e.g. mutation burden and immune gene expression). However, these markers are not sufficient to accurately predict a response to ICB therapy. Here, we analysed DNA methylomes of 141 advanced NSCLC samples subjected to ICB therapy (i.e., Anti-programmed death-1) from two independent cohorts (60 and 81 patients from our and IDIBELL cohorts). Integrative analysis of patients with matched transcriptome data in our cohort (n=28) at pathway level revealed significant overlaps between promoter hypermethylation and transcriptional repression in nonrespondersrelative to responders. Fifteen immune-related pathways, including interferon signaling, were identified to be enriched for both hypermethylation and repression. Global methylation loss correlated with immune evasion signatures independently of mutation burden and aneuploidy. Higher predictive power was observed for methylation loss than mutation burden. Hence, DNA methylation alterations implicate epigenetic modulation in precision immunotherapy. Next, we built a reliable prognostic risk model based on eight genes using LASSO model and successfully validated the model in an independent cohort. Furthermore, we found 30 survival-associated molecular interaction networks, in which two or three hypermethylated genes showed significant mutual exclusion across nonresponders. Our study demonstrates that methylation patterns can provide insight into molecular determinants underlying the clinical benefit of ICB therapy.