Objective To understand the differences in the number of drug-resistant genes, virulence genes and plasmids among carbapenem-resistant Klebsiella pneumoniae (CRKP) from different specimen sources, and provide references for the prevention and control of CRKP infection. Methods The whole-genome sequencing data of CRKP from different specimen sources in 2011-2020 were downloaded from public database of Pathsystems Resource Integration Center (PATRIC). Drug-resistant genes, virulence genes, and plasmids were identified and analyzed with bioinformatics analysis software and compared with IBM SPSS 23.0 and R v4.2.0 software. Results A total of 2 356 CRKP whole-genome data was included. The medians (interquartile interval) of drug-resistant genes, plasmids, and virulence genes carried by CRKP strains were 18 (16, 20), 4 (3, 5), and 60 (57, 68) respectively. The number of virulence genes, plasmids, and key virulence genes (iutA, iucA, iucB, iucC, iucD, rmpA, rmpA2, iroB, iroC, iroD, iroN) carried by CRKP strains from different specimen sources was statistically significant (P < 0.05). The number of drug-resistant genes, plasmids and virulence genes carried by CRKP strains from fecal specimens was higher, with the medians (interquartile interval) of 19 (16, 21), 5 (3, 6) and 63 (61, 59) respectively. The percentage of CRKP from catheter-attached fluid carrying iutA and iuc (iucA, iucB, iucC, iucD) was higher (19.0%, 4/21). Carrying rates of rmpA and rmpA2 in CRKP from alveolar lavage fluid were the highest, 7.9% (3/38) and 15.8% (6/38) respectively. The percentage of iro (iroB, iroC, iroD, iroN) genes carried by CRKP from sterile body fluids (6.5%, 4/62) was higher than that of CRKP from other specimen types (percentage range: 0-4.9%). CRKP from blood specimen carried more virulence genes than that from other specimen sources (P=0.004). Conclusion CRKP from blood specimen carries more virulence genes than CRKP from other types of specimens, which may relate to the high clinical mortality. It is suggested to pay close attention to the evolution of CRKP strains into high virulence strain in blood specimen, and reduce the incidence of CRKP blood stream infection by strengthening healthcare-associated infection control measures.