It has become increasingly clear that anti-tuberculosis regimens need optimization. Information gained using pharmacokinetics/ pharmacodynamics (PK/PD) methods in hollow fiber and animal model studies, in conjunction with Monte Carlo simulations, can be used to achieve this goal. PK/PD models of anti-tuberculosis drugs in hollow fibers, mice and guinea pigs have been remarkably concordant. Using exposures derived in these models it has been shown that the standard doses of pyrazinamide, rifampin, and ethambutol should be increased for a better efficacy, while doses of isoniazid need to be individualized. In addition, PK/PD driven doses have been proposed for new anti-tuberculosis agents such as moxifloxacin and PA-824.