Preeclampsia is a prevalent pregnancy hypertensive disease with both maternal and fetal morbidity and mortality. Emerging evidence indicates that global placental DNA hypomethylation is observed in patients with preeclampsia and is linked to altered gene expression and disease development. However, the molecular basis underlying placental epigenetic changes in preeclampsia remains unclear. Using 2 independent experimental models of preeclampsia, adenosine deaminase–deficient mice and a pathogenic autoantibody-induced mouse model of preeclampsia, we demonstrate that elevated placental adenosine not only induces hallmark features of preeclampsia but also causes placental DNA hypomethylation. The use of genetic approaches to express an adenosine deaminase minigene specifically in placentas, or adenosine deaminase enzyme replacement therapy, restored placental adenosine to normal levels, attenuated preeclampsia features, and abolished placental DNA hypomethylation in adenosine deaminase–deficient mice. Genetic deletion of CD73 (an ectonucleotidase that converts AMP to adenosine) prevented the elevation of placental adenosine in the autoantibody-induced preeclampsia mouse model and ameliorated preeclampsia features and placental DNA hypomethylation. Immunohistochemical studies revealed that elevated placental adenosine–mediated DNA hypomethylation predominantly occurs in spongiotrophoblasts and labyrinthine trophoblasts and that this effect is independent of A2B adenosine receptor activation in both preeclampsia models. Extending our mouse findings to humans, we used cultured human trophoblasts to demonstrate that adenosine functions intracellularly and induces DNA hypomethylation without A2B adenosine receptor activation. Altogether, both mouse and human studies reveal novel mechanisms underlying placental DNA hypomethylation and potential therapeutic approaches for preeclampsia.