Clinical data on the use of artesunate combined with mefloquine in a variety of treatment regimens and parasite loads in Thailand were modelled on the basis of experimentally determined pharmacokinetic data. The model assumed no pharmacodynamic interaction between artesunate and mefloquine, but that the parasites were already resistant to mefloquine. Predictions of the model accorded well with the data. In particular, in accordance with clinical observations, the model showed that monotherapy with either drug failed to cure at moderate parasitaemia, yet such patients could be treated effectively with the combination of 3 days of artesunate + mefloquine. For high levels of parasitaemia, 5 days of artesunate+mefloquine were needed. Simulations were also performed for situations of lower resistance to mefloquine and for the immune human populations found in Africa. The importance of mathematical modelling of combination therapy is borne out by this study and suggests its wider application for other drug combinations.