Encapsulating antibiotics such as rifampicin in biodegradable nanoparticles provides several advantages compared to free drug administration, including reduced dosing due to localized targeting and sustained release. Consequently, these characteristics reduce systemic drug toxicity. However, new nanoformulations need to be tested in complex biological systems to fully characterize their potential for improved drug therapy. Tuberculosis, caused by infection with the bacterium Mycobacterium tuberculosis, requires lengthy and expensive treatment, and incomplete therapy contributes to an increasing incidence of drug resistance. Recent evidence suggests that standard therapy may be improved by combining antibiotics with bacterial efflux pump inhibitors, such as thioridazine. However, this drug is difficult to use clinically due to its toxicity. Here, we encapsulated thioridazine in poly(lactic-co-glycolic) acid nanoparticles and tested them alone and in combination with rifampicin nanoparticles, or free rifampicin in macrophages and in a zebrafish model of tuberculosis. Whereas free thioridazine was highly toxic in both cells and zebrafish embryos, after encapsulation in nanoparticles no toxicity was detected. When combined with rifampicin nanoparticles, the nanoparticles loaded with thioridazine gave a modest increase in killing of both Mycobacterium bovis BCG and M. tuberculosis in macrophages. In the zebrafish, the thioridazine nanoparticles showed a significant therapeutic effect in combination with rifampicin by enhancing embryo survival and reducing mycobacterial infection. Our results show that the zebrafish embryo is a highly sensitive indicator of drug toxicity and that thioridazine nanoparticle therapy can improve the antibacterial effect of rifampicin in vivo.