Objectives Despite controversies, direct oral anticoagulants (DOACs) are increasingly used in antiphospholipid syndrome (APS). We investigated the safety and efficacy of DOACs versus vitamin K antagonists (VKAs) in real-life consecutive APS patients. Patients and methods In a cohort study of 176 APS patients, which included 82 subjects who preferred DOACs or had unstable anticoagulation with VKAs, we recorded venous thromboembolism (VTE), cerebrovascular ischemic events or myocardial infarction, along with major bleeding or clinically relevant non-major bleeding (CRNMB). Results APS patients were followed for a median time of 51 (interquartile range 43–63) months. Patients on DOACs and those on VKAs were similar with regard to baseline characteristics. APS patients treated with DOACs had increased risk of recurrent thromboembolic events and recurrent VTE alone compared with those on VKAs (hazard ratio (HR) = 3.98, 95% confidence interval (CI): 1.54–10.28, p = 0.004 and HR = 3.69, 95% CI: 1.27–10.68, p = 0.016, respectively) with no differences between rivaroxaban and apixaban or single- or double-positive and triple-positive APS. Thromboembolism on DOACs was associated with older age (median 52 versus 42 years, p = 0.008) and higher global APS score (median 13 versus 8.5, p = 0.013). Patients on DOACs had increased risk of major bleeding or CRNMB (HR = 3.63, 95% CI: 1.53–8.63, p = 0.003), but rates of gastrointestinal bleeds (HR = 3.36, 95% CI: 0.70–16.16, p = 0.13) and major bleeds or CRNMB other than heavy menstrual bleeding (HR = 2.45, 95% CI: 0.62–9.69, p = 0.2) were similar in both treatment groups. Conclusion During long-term follow-up of real-life APS patients, DOACs are less effective and less safe as VKAs in the prevention of thromboembolism.