New therapeutic strategies for rapid and effective treatment of tuberculosis are highly desirable, and their development can be drastically accelerated by facile genetic manipulation methods in Mycobacterium tuberculosis. CRISPR base editors allow for rapid, robust, and programmed single base substitutions and gene inactivation, yet no such systems are currently available in M. tuberculosis. By screening distinct CRISPR base editors, we identified that only the unusual Streptococcus thermophilus Cas9 (St1Cas9) cytidine base editor (CBE), but not the widely used Streptococcus pyogenes Cas9 or Lachnospiraceae bacterium Cpf1 CBEs, are active in mycobacteria. Despite the notable C-to-T conversions, a high portion of undesired byproducts existed with St1Cas9 CBE. We thus engineered St1Cas9 CBE by uracil DNA glycosylase inhibitor (UGI) or uracil DNA glycosylase (UNG) fusion, yielding two new base editors (CTBE and CGBE), capable of C-to-T or C-to-G conversions with dramatically enhanced editing product purity and multiplexed editing capacity. Because wild-type St1Cas9 recognizes a relatively strict PAM sequence for DNA targeting, we evolved a PAM-expanded St1Cas9 variant by structure-guided protein engineering for the base editors, substantially broadening the targeting scope. Our approaches significantly reduce the efforts and time for precise genetic manipulation and will facilitate functional genomics and drug-target exploration in M. tuberculosis and related organisms.