Until recently, vitamin K antagonists (VKAs) were the mainstay of oral anticoagulant treatment with bleeding as the most prevalent adverse effect. One to four percent of patients experience major bleeding episodes, while clinically relevant bleeding occurs in up to 20%. At this moment no laboratory assays are available to identify patients at risk for bleeding. With this study we aimed to investigate whether thrombin generation tests might identify a bleeding risk in patients taking VKAs. This prospective cohort study included 129 patients taking VKAs for more than three months. Calibrated automated thrombinography (CAT) was performed in whole blood, platelet rich and platelet poor plasma. Hematocrit, hemoglobin concentrations and the International Normalized Ratio (INR) were defined and coagulation factor levels were measured. Forty clinically relevant bleeding episodes were registered in 26 patients during follow-up. No differences were found in plasma CAT parameters or INR values. Bleeding was not associated with age, sex, hematocrit, hemoglobin levels or coagulation factor levels. In whole blood a significantly lower endogenous thrombin potential (ETP) and peak were found in patients with bleeding (median ETP: 182.5 versus 256.2 nM.min, p = 0.002; peak: 23.9 versus 39.1 nM, p = 0.029). Additionally, the area under the receiver operating curve (AUC ROC) was significantly associated with bleeding (ETP: 0.700, p = 0.002; peak: 0.642, p = 0.029). HAS-BLED scores were also significantly higher in bleeding patients (3 versus 2, p = 0.003), with an AUC ROC 0.682 (p = 0.004). In conclusion, bleeding in patients taking VKAs is associated with a decreased whole blood ETP and peak as well as with an increased HAS-BLED score.