Abstract: Carbon fibers were horizontally placed in front of the mass spectrometer inlet and aligned on the same axis as the inlet. Condom samples required no prior pretreatment, and the carbon fiber ionization mass spectrometry （CFI-MS） technique was used for rapid and direct analysis of condoms from different sources. The Xcalibur 2.2 software package was employed to process the mass spectra and output the mass spectral data, and there were significant differences in the resulting mass spectra of 60 condom samples. Partial least squares-discriminant analysis （PLS-DA） was applied for supervised pattern recognition of condom samples from Brands No.1 to No.6, and the samples were accurately classified into 6 distinct brands. Samples from No.1 to No.60 were used as the training set to establish a traceability model for condom samples, followed by necessary model optimization. Subsequently, a binary classification model was built for each brand to enable rapid and non-destructive authenticity identification of condoms, ultimately achieving traceability analysis of condom samples. CFI-MS technique had advantages of convenient operation, rapid analysis, no impurity introduction, and no damage to the test material （preservation of physical evidence）, which could provide an analytical strategy for the rapid analysis, identification, and traceability of condom physical evidence in forensic science fields such as public security investigation and judicial litigation.