This study represents a novel proof of concept of the clinical utility of miRNAs from exhaled breath condensate (EBC) as biomarkers of lung cancer (LC). Genome-wide miRNA profiling and machine learning analysis were performed on EBC from 21 healthy volunteers and 21 LC patients. The levels of 12 miRNAs were significantly altered in EBC from LC patients where a specific signature of miR-4507, miR-6777-5p and miR-451a distinguished these patients with high accuracy. Besides, a distinctive miRNA profile between LC adenocarcinoma and squamous cell carcinoma was observed, where a combined panel of miR-4529-3p, miR-8075 and miR-7704 enabling discrimination between them. EBC levels of miR-6777-5p, 6780a-5p and miR-877-5p predicted clinical outcome at 500 days. Two additional miRNA signatures were also associated with other clinical features such as stage and invasion status. Dysregulated EBC miRNAs showed potential target genes related to LC pathogenesis, including CDKN2B, PTEN, TP53, BCL2, KRAS and EGFR. We conclude that EBC miRNAs might allow the identification, stratification and monitorization of LC, which could lead to the development of precision medicine in this and other respiratory diseases.