The geometric elements of the road, such as tangents and curves, play a vital role in road safety because significant crashes are reported on the horizontal curves and tangent-to-curve transitions. Literature reveals that inconsistent geometric design of roads violates driver's expectation of operating speed leading to crashes. For safe manoeuver, it is necessary to achieve consistent operating speed with road geometry based on the driver's expectations rather than the designer's perception. Estimation of reliable operating speeds in the design phase will help to design safer road alignments. Several past studies developed operating speed models on the curves and tangent-to-curve transitions. However, these models used spot speed data with the assumption that the constant speed persists on the horizontal curves and entire deceleration/acceleration occurs on the approach/departure tangents. In this study, an instrumented vehicle with a high-end GPS (global positioning system) device was used to obtain the continuous speed profiles for passenger cars which resulted in reliable and robust peed prediction models on a tangent, curve, and tangent-to-curve. to speed prediction models for a tangent, curve, and tangent-to-curve. The study also establishes a relationship between the differential of the 85th percentile speed (Î"V85) and 85th percentile speed differential (Î"85V). The analysis results revealed that Î"V85 underestimates Î"85V by 5.32 km/h, and Î"85V predicted the actual speed reduction from tangent-to-curve transitions. Statistical analysis results showed low errors, variations, and strong correlation of the proposed models with the field data. The models developed in the present study were validated and compared with various other models available in the literature. The comparative study highlights the importance of using continuous speed profile data to calibrate the operating speed models. © 2020 American Society of Civil Engineers.