<p>This study develops time specific and vehicle class wise operating speed prediction (OSP) models on rural highway curves under both daytime and nighttime conditions. Existing studies in India largely overlook nighttime behaviour and depend on low-resolution geometric data, limiting their ability to capture real-world mixed-traffic dynamics. The objective of the present study is to analyse the influence of geometric features and temporal variations on vehicle operating speeds at critical curve locations. The study was conducted on a 30-km section of State Highway-1 (SH-1) in Karnataka, India, comprising 32 horizontal curves. High resolution LiDAR based topographic survey was used to extract geometric parameters, while more than 110,000 spot speed observations were collected with laser speed cameras at five critical locations on each curve for motorised two wheelers (MTW), passenger cars (CAR) and heavy commercial vehicles (HCV). Backward stepwise multiple linear regression was applied to develop point specific OSP models for three curve points namely, the point of curvature (PC), the midpoint (MC) and the point of tangency (PT). Results indicate that the preceding point speed is the most influential predictor, highlighting the role of geometric continuity on driver speed choice. Curve radius and degree of curvature also significantly affected vehicle speed change behaviour. The MTW and CAR based models performed better during nighttime, suggesting that reduced visibility conditions encourage more uniform and cautious driving patterns. The findings emphasize the importance of incorporating vehicle class, temporal effects and geometric continuity into highway design and speed management strategies for rural two-lane highways.</p>