© Springer International Publishing Switzerland 2016. Road traffic accidents are among the most pressing transportationrelated issues; they have not yet been addressed in a satisfactory way in many countries. They can be viewed as failures of road safety systems caused by a set of contributing components. This paper proposes a belief fault tree analysis model based on road safety inspection for identifying road infrastructure deficiencies that influence an accident occurrence and guiding highway professionals in the implementation of proper correction actions. Fault Tree Analysis is used as a risk assessment technique to diagnose the failures of road safety systems, while evidence theory is used to represent the probabilistic-based information under uncertainty gathered from expert opinions. The proposed approach is applied to analyse a real-world high-accident intersection location. It provides a means for road safety engineers to elucidate the cause of accident occurrence and to conduct road safety inspection under uncertainty.