Abstract Several case-control studies have reported elevated plasma von Willebrand factor (VWF) levels in patients with venous thromboembolism (VTE) compared with controls. However, because few studies have investigated the association in a prospective design, it is unclear whether elevated plasma VWF is a risk factor or a consequence of the VTE event. Therefore, we aimed to investigate the prospective association between plasma VWF levels and risk of VTE, as well as to perform subgroup analyses of deep vein thrombosis (DVT) and pulmonary embolism. We established a population-based nested case-control study of 414 VTE cases and 843 age- and sex-matched controls based on the Tromsø study cohort (1994-2007). Blood samples were collected at cohort baseline (1994-1995). Odds ratios (ORs) with 95% confidence intervals (CIs) for VTE were estimated across quartiles of VWF levels. We found that the risk of VTE increased linearly across quartiles of VWF levels (P for trend = .023). Participants with VWF in the highest quartile had an OR of 1.45 (95% CI, 1.03-2.03) for VTE compared with those in the lowest quartile. The association was strongest for unprovoked VTE (OR, 2.74; 95% CI, 1.66-4.54) and unprovoked DVT in particular (OR, 6.73; 95% CI, 3.07-14.76). Further adjustment for body mass index, C-reactive protein, hypertension, estrogen use, and smoking had a modest effect on the risk estimates. To conclude, we found a dose-dependent relationship between plasma VWF levels and future risk of incident VTE, and unprovoked events in particular. Our findings suggest that VWF may represent a promising biomarker for future risk of incident VTE.