This study examines road safety as a part of social and economic costs in sustainable development management. The research focuses on the road network in Warmia-Masuria Voivodeship, Poland, analyzing 180 homogeneous sections of regional roads totaling over 1,800 km. Two methods are employed: reactive and proactive. The reactive method assesses risk based on accident density and accident cost density, classifying road sections into five risk levels. Results show that 4% of the analysed network does not meet sustainable development criteria, requiring immediate action to reduce risk. The proactive method utilizes regression models to predict social losses, expressed as pedestrian accident density, injured victim density, and severely injured and fatality density. These models consider factors such as traffic volume and road characteristics. The study demonstrates that road safety measures can effectively evaluate network sustainability from both social and economic perspectives, allowing for targeted interventions to improve safety and sustainability in road infrastructure management.