The major histologic subtypes of non-small cell lung cancer (NSCLC) include adenocarcinoma (ADC), squamous cell lung carcinoma (SCC), and large-cell carcinoma (LCC). Clinical trials of targeted agents and newer chemotherapy agents yielded differences in outcomes according to histologic subgroups providing a rationale for histology-based treatment in NSCLC. Currently, NSCLC subtyping is performed based on histopathological examinations and immunohistochemistry. However available methods leave about 10% of NSCLC cases as not otherwise specified. The purpose of this study was development of an LC-QTOF-MS method for human lung tissue metabolic fingerprinting that could discriminate NSCLC histological subtypes and propose biomarkers candidates that could support proper NSCLC diagnosis. Metabolites were extracted with acetonitrile or methanol/ethanol and different chromatographic conditions were tested. In the final method 10 mg of lung tissue was homogenized with 50% methanol and metabolites were extracted with acetonitrile. Metabolites were separated on C8-RP and HILIC columns. About 3500 and 2000 of metabolic features (in both ion modes) were detected with good repeatability (CV < 20%) by RP and HILIC methods, respectively. Lung tumor and control tissue samples obtained from NSCLC patients were analyzed with developed methodology. Acylcarnitines, fatty acids, phospholipids, and amino acids were found more abundant in tumor as compared to control tissue. Acylcarnitines, lysophospholipids, creatinine, creatine, and alanine were identified as potential targets enabling classification of NSCLC subtypes.