This study aims to evaluate the efficacy of anti-factor Xa activity (aFXa) in predicting ecchymosis after total knee arthroplasty (TKA). One hundred and two unilateral primary TKA patients were recruited consecutively in this prospective observational study. Participants received rivaroxaban (10 mg p.o. qd) from postoperative day 1 (POD1) to POD35 and were divided into a non-ecchymosis group (group A) and an ecchymosis group (group B). AFXa was assessed as the primary outcome on POD1 and POD3. Prothrombin time (PT), activated partial thromboplastin time (APTT) and thromboelastography (TEG) were recorded both preoperatively and postoperatively (on POD1 and POD3). Other outcomes, including venous thromboembolism (VTE), blood loss and wound complications were also collected and compared. As a result, 27.5% of the participants (n = 28) were allocated into group B. Demographic data were comparable between the two groups. The aFXa levels in group B were significantly higher than those in group A on POD1 and POD3, and the aFXa level was assessed as an independent risk factor for ecchymosis. The cut-off value of aFXa was determined to be 121.38 ng/mL at maximal Youden index, associated with area under the receiver operating characteristics curve of 0.67. Group B experienced significantly more blood loss and wound complications than group A. No statistical difference was detected regarding PT, APTT and TEG parameters. AFXa is a promising parameter to predict ecchymosis after TKA. Patients with aFXa > 121.38 ng/mL should be considered as high-risk population for postoperative ecchymosis and may require intense monitoring or dosage modification of anticoagulants.