Tuberculosis (TB), caused by the intracellular bacterium <i>Mycobacterium tuberculosis</i> (Mtb), remains a major health threat. A live, attenuated mycobacterium known as Bacille Calmette-Guérin (BCG), derived from the causative agent of cattle TB, <i>Mycobacterium bovis</i>, has been in clinical use as a vaccine for 90 years. The current incidence of TB demonstrates that BCG fails to protect sufficiently against pulmonary TB, the major disease manifestation and source of dissemination. The protective efficacy of BCG is on average 50% but varies substantially with geographical location and is poorer in those with previous exposure to mycobacteria. BCG can also cause adverse reactions in immunocompromised individuals. However, BCG has contributed to reduced infant TB mortality by protecting against extrapulmonary TB. In addition, BCG has been associated with reduced general childhood mortality by stimulating immune responses. In order to improve the efficacy of BCG, two major strategies have been employed. The first involves the development of recombinant live mycobacterial vaccines with improved efficacy and safety. The second strategy is to boost BCG with subunit vaccines containing Mtb antigens. This article reviews recombinant BCG strains that have been tested against TB in animal models. This includes BCG strains that have been engineered to induce increased immune responses by the insertion of genes for Mtb antigens, mammalian cytokines, or host resistance factors, the insertion of bacterial toxin-derived adjuvants, and the manipulation of bacterial genes in order to increase antigen presentation and immune activation. Subunit vaccines for boosting BCG are also briefly discussed.