Road markings are part of traffic signalization that form the road surface by using and combining lines, signs and symbols and provide information concerning visual guidance to road users. Since they are positioned in the drivers’ central field of vision, road markings are an important element contributing to road safety. In order to perform their function and contribute to the increase of road safety, road markings have to be visible, especially under low visibility conditions. It is precisely for these reasons that road markings visibility represents their most important characteristic for road safety, based on which the cycle of maintenance or renewal of road markings is determined. In this doctoral thesis models have been developed to predict service life of road markings for three most frequently used materials: solventborne paint, thermoplastic and cold structural plastic. The research is based on the data obtained by measuring visibility, that is, retroreflection of road markings by means of a contemporary dynamic method for measuring road markings retroreflection. In addition to retroreflection and in order to obtain accurate and precise models, based on the existing literature, relevant factors have been identified that affect the degradation of road markings retroreflection and thus their service life. Such factors include: road marking age and position, average annual daily traffic, winter maintenance activities and average speed limit on the road. Their impact on the road markings service life has been statistically analysed in detail. The research comprised in total 115 road markings applied in solventborne paint in total length of 5 218 km, 30 thermoplastic road markings in total length of 579 km and 30 cold structural plastic road markings in total length of 497,9 km. Models for predicting the service life of road markings have been developed based on the statistical analysis of the impact of individual factors on the durability of the stated materials used for application of road markings. The developed models have shown a satisfactory accuracy and as such, they can be efficiently used to optimize road maintenance activities, ultimately aimed at increasing road safety and rationalize maintenance costs.