Background: Exhaled breath-based test is an attractive option for cancer detection due to its non-invasive nature. Exhaled volatile organic compounds (VOCs) are produced in various biochemical processes and might be sensitive tumor biomarkers. Here, we reported an exploratory study to investigate the performance of exhaled VOCs for detection of early-stage lung cancer using a high-resolution high-pressure photon ionization time-of-flight mass spectrometry (HPPI-TOFMS). Methods: Treatment-naïve patients with pulmonary nodules who received surgery at our department and without history of cancer were enrolled. Exhaled breath samples were collected before surgery and stored in Tedla bags. A CO₂ sensor was applicated during sample collection to ensure only “alveolar air” was collected. Exhaled samples were directly detected by HPPI-TOFMS, which has a resolution > 3000. Deep learning algorithm was used to build detection model based on HPPI-TOFMS data. Results: A total of 171 patients were included in this study, including 139 patients with lung cancer (114 of TNM stage I, 14 of stage II, 9 of stage III, and 2 of stage IV) and 32 patients with benign nodules. Mass spectrum peaks with m/z< 500 detected by HPPI-TOFMS were retained and 32500 features were extracted from each exhaled breath samples. Based these extracted features, participants who were pathologically diagnosed as lung cancer could be discriminated from those with benign diseases with an accuracy of 96.19%, sensitivity of 96.43%, and specificity of 84.38%. Discrimination of lung cancer patients with lymph node metastasis (n = 12) from those without lymph node metastasis (n = 127) had an accuracy of 83.23%. Conclusions: Exhaled VOCs as detected by a high-resolution HPPI-TOFMS might be sensitive biomarkers for detection of early-stage lung cancer.