Metabolic fingerprinting, the main tool in metabolomics, is a non-targeted methodology where all detectable peaks (or signals), including those from unknown analytes, are considered to establish sample classification. After pattern comparison, those signals changing in response to a specific situation under investigation are identified to gain biological insight. For this purpose, gas chromatographymass spectrometry (GC-MS) has a drawback in that only volatile compounds or compounds that can be made volatile after derivatization can be analysed, and derivatization often requires extensive sample treatment. However, once the analysis is focused on low molecular weight metabolites, GC-MS is highly efficient, sensitive, and reproducible. Moreover, it is quantitative, and its compound identification capabilities are superior to other separation techniques because GC-MS instruments obtain mass spectra with reproducible fragmentation patterns, which allow for the creation of public databases. This chapter describes well-established protocols for metabolic fingerprinting (i.e. the comprehensive analysis of small molecules) in plasma and urine using GC-MS. Guidelines will also be provided regarding subsequent data pre-treatment, pattern recognition, and marker identification.