Tuberculosis remains a major cause of mortality and physical and economic deprivation worldwide. There have been significant recent advances in our understanding of the Mycobacterium tuberculosis genome, mycobacterial genetics and the host determinants of protective immunity. Nevertheless, the challenge is to harness this information to develop a more effective vaccine than BCG, the attenuated strain of Mycobacterium bovis derived by Calmette and Guérin nearly 90 years ago. Some of the limitations of BCG include the waning of the protective immunity with time, reduced effectiveness against pulmonary tuberculosis compared to disseminated disease, and the problems of a live vaccine in immuno-compromised subjects. Two broad approaches to vaccine development are being pursued. New live vaccines include either attenuated strains of Mycobacterium tuberculosis produced by random mutagenesis or targeted deletion of putative virulence factors, or by genetic manipulation of BCG to express new antigens or cytokines. The second approach utilizes non-viable subunit vaccines to deliver immunodominant mycobacterial antigens. Both protein and DNA vaccines induce partial protection against experimental tuberculosis infection in mice, however, their efficacy has generally been equivalent to or less than that of BCG. The comparative effects of cytokine adjuvants and vaccines targeting antigen presenting cells on enhancing protection will be discussed. Coimmunization with plasmid interleukin-12 and a DNA vaccine expressing Antigen 85B, a major secreted protein, was as protective as BCG. The combination of priming with DNA-85B and boosting with BCG was superior to BCG alone. Therefore it is possible to achieve a greater level of protection against tuberculosis than with BCG, and this highlights the potential for new tuberculosis vaccines in humans.