Preclinical studies in rodents have demonstrated that artemisinins, especially injectable artesunate, can induce fetal death and congenital malformations at a low dose range. The embryotoxicity can be induced in those animals only within a narrow window in early embryogenesis. Evidence was presented that the mechanism by which embryotoxicity of artemisinins occurs seems to be limited to fetal erythropoiesis and vasculogenesis/ angiogenesis on the very earliest developing red blood cells, causing severe anemia in the embryos with higher drug peak concentrations. However, this embryotoxicity has not been convincingly observed in clinical trials from 1,837 pregnant women, including 176 patients in the first trimester exposed to an artemisinin agent or artemisinin-based combination therapy (ACT) from 1989 to 2009. In the rodent, the sensitive early red cells are produced synchronously over one day with single or multiple exposures to the drug can result in a high proportion of cell deaths. In contrast, primates required a longer period of treatment of 12 days to induce such embryonic loss. In humans only limited information is available about this stage of red cell development; however, it is known to take place over a longer time period, and it may well be that a limited period of treatment of 2 to 3 days for malaria would not produce serious toxic effects. In addition, current oral intake, the most commonly used route of administration in pregnant women with an ACT, results in lower peak concentration and shorter exposure time of artemisinins that demonstrated that such a concentration-course profile is unlikely to induce the embryotoxicity. When relating the animal and human toxicity of artemisinins, the different drug sensitive period and pharmacokinetic profiles as reviewed in the present report may provide a great margin of safety in the pregnant women.