Background: The chloroplast (cp) genome has unique and highly conserved characteristics and is therefore widely used in species identification and classification, as well as to improve the in–depth understanding of plant evolution. Methods: In this study, the cp genomes of 13 Lamiaceae plants in the Tibet Autonomous Region of China were sequenced, assembled and annotated using bioinformatics methods. Phylogenetic trees were constructed to reveal the phylogenetic relationship of related species in the Lamiaceae. Results: The results showed that all 13 cp genomes had a typical four–segment structure, including one large single–copy (LSC) region, one pair of inverted repeat (IR) regions and one small single–copy (SSC) region. The sequence lengths of the 13 cp genomes were between 149,081 bp and 152,312 bp, and the average GC content was 37.6%. These genomes contained 131–133 annotated genes, including 86–88 protein–coding genes, 37–38 tRNA genes, and 8 rRNA genes. A total of 542 SSR loci were detected using MISA software. The repeat types were mostly single–nucleotide repeats, accounting for 61% of simple repeats. A total of 26,328–26,887 codons were detected in 13 cp genomes. According to the RSCU value analysis, the codons mostly ended with A/T. Analysis of IR boundaries showed that the other species were relatively conserved, except for Nepeta laevigata (D. Don) Hand.–Mazz., which differed in gene type and location on both sides of the boundary. By analysing nucleotide diversity, two highly mutated regions located in the LSC and SSC regions were identified in the 13 cp genomes. Conclusions: Using the cp genome of Lycium ruthenicum Murray as the outgroup, 97 cp genomes of the Lamiaceae were used to construct an Maximum Likehood (ML) phylogenetic tree, in which these species were divided into eight major clades, corresponding to eight subfamilies based on morphological classification. The phylogenetic results based on monophyletic relationships were consistent with the morphological classification status at the tribe level.