Anticoagulants and anti-platelet agents are simultaneously administrated in clinical practice (i.e. percutaneous coronary intervention), which cause significant risk of systemic bleeding. Targeted delivery of anticoagulants to the activated platelets at sites of vascular injuries may condense the site-specific anticoagulant effect and reduce the hemorrhage side effects in uninjured vessels. To this end, we prepared three ancylostoma caninum anticoagulant peptide 5 (AcAP5) variants NR1, NR2 and NR3 engineered with a platelet-binding Arg-Gly-Asp (RGD) motif and evaluated their anti-Factor Xa (FXa) and platelet-binding effects. These RGD-containing AcAP5 variants were capable of interacting with platelet receptor αIIbβ3 as shown in computational analysis. All variants, especially NR2 and NR3, retained entirely the anti-FXa function of parent AcAP5. Moreover, they prevented the formation of occlusive thrombi in rat carotid artery injury model, suggesting that they inhibit platelet aggregation in vivo. Further functional investigation of NR3 demonstrated that NR3 inhibited platelet aggregation in vitro and FXa activity in vivo, and prolonged the coagulation time, all in a dose-dependent manner. Through flow cytometry assay, we confirmed the binding of NR3 to αIIbβ3 receptor. In mouse model of carotid artery endothelium injury, NR3-treated mice showed less tail bleeding time than AcAP5-treated mice, and aspirin plus NR3 treatment exhibited moderate reduction of blood loss compared with aspirin plus AcAP5 treatment. These results indicate the feasibility to engineer a novel FXa inhibitor specifically targeting the activated platelets, which centralizes its anticoagulation efficacy in the injured vascular endothelium and reduces the risk of systemic bleeding.