Optimal diagnosis and management of children aged <15 y with rifampicin- or multidrug-resistant tuberculosis (RR/MDR-TB) relies on identification of adults with the disease and pro-active screening of their close contacts. Children may be diagnosed with RR/MDR-TB based on microbiological confirmation from clinical specimens (sputum, gastric washings, stool), but usually the diagnosis is presumptive, with a history of exposure to RR/MDR-TB and clinical/radiological signs and symptoms suggestive of TB disease. RR/MDR-TB should also be considered in children where first-line TB treatment fails despite good adherence to therapy. Composition and duration of all-oral RR/MDR-TB treatment regimens in children are based on site and severity of TB disease, drug resistance profile of the Mycobacterium tuberculosis strain (isolated from the child or from the most likely source patient), inclusion of at least four drugs considered to be effective (with priority given to World Health Organization Group A and B drugs), toxicity and tolerability of medications (and feasibility of adverse effect monitoring in the child’s setting), and availability of child-friendly formulations of TB medications. Individualized RR/MDR-TB regimens are preferable to the standardised 9–12-mo regimen for children, and injectable agents must not be used. Optimal adherence to treatment relies on education, training and support for caregivers and others who are responsible for administering medications to children, as well as close clinical monitoring and early management of adverse effects. Children who are initiated on adequate RR/MDR-TB regimens have high treatment success rates, but efforts to find and treat more children with undiagnosed RR/MDR-TB are crucial to reduce childhood TB mortality.