Thalidomide, which was formerly available commercially to control the symptoms of morning sickness, is a strong teratogen that causes fetal abnormalities. However, the mechanism of thalidomide teratogenicity is not fully understood; thalidomide toxicity is not apparent in rodents, and the use of human embryos is ethically and technically untenable. In this study, we designed an experimental system featuring human-induced pluripotent stem cells (hiPSCs) to investigate the effects of thalidomide. These cells exhibit the same characteristics as those of epiblasts originating from implanted fertilized ova, which give rise to the fetus. Therefore, theoretically, thalidomide exposure during hiPSC differentiation is equivalent to that in the human fetus. We examined the effects of thalidomide on undifferentiated hiPSCs and early-differentiated hiPSCs cultured in media containing bone morphogenetic protein-4, which correspond, respectively, to epiblast (future fetus) and trophoblast (future extra-embryonic tissue). We found that only the number of undifferentiated cells was reduced. In undifferentiated cells, application of thalidomide increased the number of apoptotic and dead cells at day 2 but not day 4. Application of thalidomide did not affect the cell cycle. Furthermore, immunostaining and flow cytometric analysis revealed that thalidomide exposure had no effect on the expression of specific markers of undifferentiated and early trophectodermal differentiated cells. These results suggest that the effect of thalidomide was successfully detected in our experimental system and that thalidomide eliminated a subpopulation of undifferentiated hiPSCs. This study may help to elucidate the mechanisms underlying thalidomide teratogenicity and reveal potential strategies for safely prescribing this drug to pregnant women.