In Bosnia and Herzegovina (B&H), two-lane rural highways make up the largest percentage of road network and accommodate almost complete traffic. Because of this, platoons are often formed and vehicle speeds are reduced. To improve the quality of traffic flow, it is necessary to define the level of service (LOS) for the existing condition and for alternative possibilities of reconstructing critical elements of the road network. According to Highway Capacity Manual 2010 (HCM 2010), for two-lane highways LOS is defined by measures of efficiency: average travel speed (ATS) for road class I and percent time spent following (PTSF) for road classes I and II. HCM assumes very high ATS values for conditions on national roads that are classified as class I according to HCM classification, and the problem is also nonuniformity of highway geometry elements, which further complicates the application of ATS in B&H. Therefore, PTSF was studied in this paper as an appropriate measure for determining the efficiency of two-lane rural highways. Researches carried out in the world show significant differences in PTSF results compared to the HCM calculations. No research on this topic has been conducted in the region and HCM is used in its original form. The above reasons were the main motive to study this topic and develop a model that would be easy to apply in practice. The survey was conducted on main roads with calculation speed of 80 km/h, because they are the most important in the two-lane road network. In this dissertation, calibration and validation of the Vissim 9 simulation software was first made based on conducted field research. The reason for application of the simulation software is that it allows simulation of different traffic conditions that occur in the field but cannot all be covered by surveying, or for testing some new traffic conditions. Based on the simulations of different traffic conditions, a new model for calculation of PTSF on two-lane rural roads (for speed 80 km/h) was obtained. The obtained PTSF calculation model was validated on another section and the obtained results proved the applicability of the model.