Abstract: Horizontal curves are often associated with roadway crashes due to speed misjudgment and loss of control. With growing adoption of autonomous and connected vehicles, accurate safe curve speed estimation is increasingly important. The widely used AASHTO design method relies on a simplified point mass model using conservative parameters to account for vehicular and environmental variations. This paper presents a digital twin-based framework for estimating safe curve speeds using a physics-driven Unity environment. A real-world horizontal curve is selected and speed data collected via radar gun under various weather conditions. A 3D road model is constructed using geometric and elevation data, with a parameterized vehicle model allowing variations in mass, acceleration, and center of gravity to reflect different vehicle types and loading scenarios. The simulation identifies the maximum safe traversal speed, offering a more vehicle- and environment-specific estimate. Estimates were validated against real-world observed speeds. This study demonstrates how a physics-based digital twin can deliver safer, more adaptive speed estimates for vehicles traversing horizontal curves.