Managing safety risk on construction projects represents the mainstay for projects’ success. In fact, by avoiding accidents, all the related direct and indirect costs and delays can be prevented.  However, because of construction projects inherent characteristics, many unplanned events may occur. Therefore, understanding accidents occurrence comportment seems to be the most important Construction Management challenge.

Different methods have been developed for accidents risk prediction such as Risk Assessment, but they are always confronted to the reality of the construction industry. In addition, these methods are not systematically taking into consideration human behavior factor, which is far from being modelizable by simple linear mathematic models. As an alternative, a concept that combines Fault Tree Analysis and Task Analysis is proposed in this study. This method considers construction site as a complex system in which failures may occur. The root causes of these failures will draw Fault Trees that can be used for probabilistic qualitative and quantitative simulations. The aim is to develop a customizable decision support tool able not only to alarm against the risks of accidents, but also to give recommendations for actions that can be implemented in order to minimize accidents occurrence probability.