Low molecular weight heparin (LMWH) is being tested as an experimental drug for improving pregnancy outcome in women with inherited thrombophilia and placenta-mediated pregnancy complications, such as recurrent pregnancy loss. The role of thrombotic processes in these disorders remains unproven, and the issue of antithrombotic prophylaxis is intensely debated. Using a murine model of factor V Leiden-associated placental failure, we show that treatment of the mother with LMWH allows placental development to proceed and affords significant protection from fetal loss. Nonetheless, the therapeutic effect of LMWH is not replicated by anticoagulation; fondaparinux and a direct Xa inhibitor, C921-78, achieve anticoagulation similar to LMWH but produce little or no improvement in pregnancy outcome. Genetic attenuation of maternal platelet aggregation is similarly ineffective. In contrast, even a partial loss of thrombin sensitivity of maternal platelets protects pregnancies. Neonates born from these pregnancies are growth retarded, suggesting that placental function is only partially restored. The placentae are smaller but do not reveal any evidence of thrombosis. Our data demonstrate an anticoagulation-independent role of LMWH in protecting pregnancies and provide evidence against the involvement of thrombotic processes in thrombophilia-associated placental failure. Importantly, thrombin-mediated maternal platelet activation remains central in the mechanism of placental failure.