Lyme disease, caused by Borrelia burgdorferi, is a focally endemic tick‐transmitted zoonotic infection. In this study, the major factors underlying synonymous codon‐related amino acid usage in the B. burgdorferi genome and bias in synonymous codon usage of the translation initiation region of coding sequences were analyzed. Additionally, adaptation of B. burgdorferi to several of its hosts was analyzed in the context of synonymous codon usage. Principal component analysis (PCA) revealed that nucleotide content at the third synonymous position of a codon influenced the synonymous codon usage pattern, but the strand‐specific factor did not influence the synonymous codon usage pattern of B. burgdorferi. In terms of the low GC content of B. burgdorferi coding sequences, the effective number of codons (ENC) showed a significant correlation with GC3 content (at the synonymous position). For the amino acid usage pattern for B. burgdorferi, PCA showed that the strand‐specific factor did not contribute to this pattern, while the properties (aromaticity and hydrophobicity) of the amino acids themselves showed strong correlations with this pattern. Under‐represented codons, which were frequently selected in the translation initiation region, possibly play roles in regulating gene expression in B. burgdorferi. In terms of co‐evolution and synonymous codon usage patterns, adaptation of B. burgdorferi to different intermediate hosts was apparent to different degrees, and the degree of adaptation of this spirochete to wild animals was stronger than that of humans or mice.