Curves are the leading cause of road departure crashes that turn into deaths and severe injuries. Most crashes occur on curved roads because they have different geometrical circumstances from straight roads. A sharp turn on the road creates more difficulty for the driver than a normal curve road. To Avoid crashes on the curved road, there is a prerequisite to identify the curved road for drivers that this turn is problematic. Responding intelligently to save precious human life and the country’s gross domestic product is necessary. Our paper proposed a Curve Crash Avoidance Protocol (2CAP) in Vehicular Ad-hoc Network (VANET). The Intelligent Curve Crash Avoidance Algorithm is proposed to avoid disastrous vehicle crashes. The sensing Node Operation algorithm works for multiple sensors embedded in the Onboard Unit to gather vehicle and road environment information. On the other side, road-side unit requests for this collected data via a secure communication channel for processing. A Linear Regression machine learning technique implements the intelligent Unit Operations algorithm. The Intelligent Unit decides to notify the onboard Unit based on gathered data and trained dataset. We implement the scheme using the Linear Regression Machine Learning Model. Multiple sensors, Global Positioning Systems, and Global Information systems use the dataset methodology to classify and predict results. The proposed model is expected to be effective for proper coordination with wireless sensor network equipment.