Abstract: OBJECTIVE To establish the rapid identification model for carbapenem-resistant Klebsiella pneumoniae (CRKP) carrying KPC-2 gene based on matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) combined with machine learning algorithm. METHODS The Klebsiella pneumoniae strains were collected from the patients who were hospitalized in Guangdong Provincial Second Hospital of Traditional Chinese Medicine from Jan. 2023 to Dec. 2023. The CRKP strains carrying KPC-2 gene were obtained by screening and verification with carbapenemase inhibitor enhancement test and PCR. The identification model for CRKP and carbapenem-sensitive Klebsiella pneumoniae(CSKP) strains carrying KPC-2 gene were established based on characteristic peak analysis by using EX-Smartspec software matching with mass spectrum and convolutional neural network algorithm, and the model was verified. RESULTS Totally 110 strains of CRKP strains carrying with KPC-2 were obtained through screening and verification. Two characteristic peaks of the CRKP strains carrying with KPC-2 gene were discovered at 4 438.1 m/z and 6 151.3 m/z through matrix analysis of protein polymerization peaks of MS, and 1 characteristic peak of CSKP was discovered at 7 317.7m/z, the characteristic peaks served as the markers for identification of the CRKP and CSKP strains carrying with KPC-2 gene. The model established based on typing training showed high accuracies (about 0.97 and 0.98) and low loss value (about 0.05 and 0.04) in the training phase and the internal verification. The external verification was carried out for 15 strains of K. pneumoniae except for the modeling, and the accuracy of identification of the CRKP and CSKP strains carrying with KPC-2 gene was 100.00%. CONCLUSION The rapid identification model for the CRKP strains carrying with KPC-2 gene that is established based on MALDI-TOF MS combined with machine learning algorithm can provide bases for rapid, reasonable treatment of CRKP and control of hospital-associated infections.