As budgets for road safety measures, policies and interventions are limited, road safety stakeholders worldwide have to decide about the most effective use of available funds. Policy makers need prediction tools and decision support systems, developed through the utilization of quantitative road safety modelling research, allowing them to analyze potential safety issues, identify appropriate safety improvements and estimate their potential effect in terms of crash reduction. Two distinct approaches can be defined in road safety modelling research, as far as the nature of the developed models and results is concerned: the microscopic (i.e. accident prediction at specific sites) and the macroscopic approach (accidents and casualty forecasts over larger geographic regions). The paper aims to present a systematic review of existing knowledge and international pertinent research on both microscopic and macroscopic road safety modelling, pointing out how they actually complement each other to the benefit of road safety practitioners.