The thermal desorption/pyrolysis-direct analysis in real time-mass spectrometry (TD/Py-DART-MS) method was developed for the analysis of fibers in this study. The fiber samples were pyrolyzed with a temperature gradient and the pyrolysis products were determined by DART-MS. The pyrogram from the TD/Py-DART-MS fiber analysis was found to be associated with the physical properties such as the melting points. At the same time, the TD/Py-DART-MS allows the analyst to obtain the chemical information such as polymeric backbone structures and dyes on the fiber. The pyrolysis profiles of common polymeric fibers in textile materials such as cotton, cellulose triacetate (CT), poly(caprolactam) (nylon-6), poly(hexamethylene adipamide) (nylon-6/6), poly(acrylonitrile) (PAN), poly(ethylene terephthalate) (PET), poly(butylene terephthalate) (PBT), poly(propylene) (PP), and polytrimethylene terephthalate (PTT) and their respective characteristic mass spectra were reported in this study. The fibers from 40 commercial textile samples were analyzed by the TD/Py-DART-MS method, and the statistical methods including principal component analysis (PCA) and Pearson product moment correlation (PPMC) were applied to classify and associate the fibers based on their mass spectral data. The strong correlation between the reference fiber mass spectral profiles and tested fiber mass spectral profiles was observed by using the PPMC method, and the identification accuracy was 97.5%. When combined, the mass spectral and pyrogram data, the types of fibers including the blended fibers were identified effectively. The TD/Py-DART-MS method also demonstrated the promising capability for the identification of dyes on fibers. Overall, the TD/Py-DART-MS method requires small sample size and minimal sample preparation but offers reproducible and multidimensional information for the fiber evidence rapidly (i.e., ∼6.7 min). Since the proposed method is simple to perform and the data are easy to interpret, this approach may significantly contribute to the fiber identification and comparison procedures in forensic settings with high sample throughput potential.