Recurrent venous thromboembolism (VTE) despite curative anticoagulation is frequent in patients with cancer. Identifying patients with a high risk of recurrence could have therapeutic implications. A prospective study was designed to validate the Ottawa risk score of recurrent VTE in cancer patients. In a prospective multicenter observational cohort, adult cancer patients with a recent diagnosis of symptomatic or incidental lower limb deep vein thrombosis or pulmonary embolism (PE) were treated with tinzaparin for 6 months. The primary endpoint was the recurrence of symptomatic or asymptomatic VTE within the first 6 months of treatment. All clinical events were centrally reviewed and adjudicated. Time-to-event outcomes were estimated by the Kalbfleisch and Prentice method to take into account the competing risk of death. A C-statistic value of > 0.70 was needed to validate the Ottawa score. A total of 409 patients were included and analyzed on an intention-to-treat basis. Median age was 68 years, 60.4% of patients had PE, and VTE was symptomatic in 271 patients (66.3%). The main primary sites were lung (31.3%), lower digestive tract (14.4%), and breast (13.9%) cancers. The Ottawa score was high (≥ 1) in 58% of patients. The 6-month cumulative incidence of recurrent VTE was 7.3% (95% confidence interval [CI]: 4.9-11.1) overall, and 5.0% (95% CI: 2.3-10.8) versus 9.1% (95%CI: 6.1-13.6) in the Ottawa low versus high risk groups, respectively. The C-statistic value was 0.60 (95% CI: 0.55-0.65). In this prospective cohort of patients with cancer receiving tinzaparin for VTE, the Ottawa score failed to accurately predict recurrent VTE.