Listeria monocytogenes is a foodborne pathogen worldwide. Multilocus variable-number tandem repeat analysis (MLVA) has been used for listeriosis surveillance and outbreak investigations. MLVA typing schemes have been proposed, but their usefulness for typing isolates from the People's Republic of China has not been assessed. To this aim, all L. monocytogenes strains (79) isolated from 1,445 raw meat and abattoir environmental samples of three western provinces in China were characterized with PCR serogrouping, multilocus sequence typing, and MLVA. The isolates were typed into the four PCR serogroups IIb (38.0%), IIc (26.6%), IIa (24.0%), and IVb (11.4%), with a Simpson's index (SI) of 0.7235. With multilocus sequence typing, they were typed into 18 sequence types (STs), including two new STs, ST1029 and ST1011, with an SI of 0.8880. With the 14 MLVA loci from the previous five schemes, the isolates were typed into 39 MLVA genotypes, with an SI of 0.9656. The typing data indicated that MLVA had the highest typing capability among the three methods. A subsequent optimization analysis identified an optimal combination of eight loci (LMV2, LMV9, LMV1, Lm10, Lm11, Lm15, Lm23, and LMTR6) producing the same SI as that of the 14 loci. The present optimized combination shared only six loci with the optimal nine-loci combination proposed in Australia, verifying for the first time that the optimal combinations varied with the isolates' sets. The current optimal typing scheme was ideal for L. monocytogenes isolates from western China.