[Objective] At present, studies on terpenoid synthase mainly focus on plants and fungi. The systematic study of bacterial terpenoid synthase is still rare. Based on a large number of bacterial genomes, we identified bacterial terpenoid synthase in genome-wide and predicted their function by bioinformatics. [Methods] Local bacterial proteome database was first built based on the bacterial protein sequences, and hidden markov model (HMM) of PF03936 in Pfam was used to predict bacterial terpenoid synthase. Then multiple sequence alignment for candidate terpenoid synthase was aligned by the tool of MAFFT 7.130b, and phylogenetic tree was constructed by MEGA 6.0. Finally, analysis of motifs was conducted by MEME and point mutation of TPS by PredictProtein. [Results] In total 1 423 terpenoid synthases were identified; they are distributed in 8 phylum, including: Actinobacteria, Proteobacteria, Cyanobacteria, Bacteroidetes, Firmicutes, Chloroflexi, Acidobacteria and Chlamydiae. Phylogenetic analysis revealed that bacterial terpenoid synthases can be divided into 4 groups. Motifs analysis of each group by MEME showed that in addition to conserved motifs among groups, there were some specific motifs within each group, which implies functional differentiation among different groups. Point mutations analysis showed that mutations of amino acid sites in different positions of terpenoid synthase could have different effect on TPS function. [Conclusion] Bacterial terpenoid synthases are mainly distributed in 8 phylum, of which 2 phylum have not been reported before; they are Firmicutes and Acidobacteria. Phylogenetic analysis showed that the difference among 4 groups of terpenoid synthase is mainly caused by group-specific motifs. In addition, mutation analysis of amino acids in different positions of terpenoid synthase provides a fundament for further verification of bacterial terpenoid synthase function.