Purpose: This paper describes a new estimation method of the number of road injuries in The Netherlands.Methods: The bases for this method are the hospital inpatient registry and the police crash record database. Both databases contain errors and omissions. The police database in particular suffers from serious underreporting, and is also inaccurate in indicating injury severity. The hospital database is inaccurate in indicating that a patient was involved in a road crash. Nonetheless, in principle it contains all serious road injuries. After linking both databases an estimating method, inspired by capture–recapture, was used to estimate the number of road injuries. The differences in registration for transport mode, injury severity, and region of crash have been taken into account.Results: This leads to an estimation of the number of serious road injuries in the Netherlands in 1993–2008.Conclusions: We found that 85% of the road injuries are recognizable as such in the hospital registry. The registration rate of the police registry is different for road injuries in crashes involving motorized vehicles (58% in 2008) and for road injuries in crashes not involving motorized vehicles (4% in 2008).