Currently available anticoagulants are effective in reducing the recurrence rate of venous thromboembolism (VTE). However, anticoagulant treatment is associated with an increased risk for bleeding complications. Thus, anticoagulation has to be discontinued when benefit of treatment no longer clearly outweigh its risks. The duration of anticoagulant treatment is currently framed based on the estimated individual risk for recurrent VTE. The incidence of recurrent VTE can be estimated through a two-step decision algorithm. Firstly, the features of the patient (gender), of the initial event (proximal or distal deep vein thrombosis or pulmonary embolism), and the associated conditions (cancer, surgery, etc) provide essential information on the risk for recurrence after anticoagulant treatment discontinuation. Secondly, at time of anticoagulant treatment discontinuation, D-dimer levels and residual thrombosis have been indicated as predictors of recurrent VTE. Current evidence suggests that the risk of recurrence after stopping therapy is largely determined by whether the acute episode of VTE has been effectively treated and by the patient's intrinsic risk of having a new episode of VTE. All patients with acute VTE should receive oral anticoagulant treatment for three months. At the end of this treatment period, physicians should decide for withdrawal or indefinite anticoagulation. Based on intrinsic patient's risk for recurrent VTE and for bleeding complications and on patient preference, selected patients could be allocated to indefinite treatment with VKA with scheduled periodic re-assessment of the benefit from extending anticoagulation. Alternative strategies for secondary prevention of VTE to be used after conventional anticoagulation are currently under evaluation. Cancer patients should receive low molecular-weight heparin over warfarin in the long-term treatment of VTE. These patients should be considered for extended anticoagulation at least until resolution of underlying disease.