Vehicular ad hoc networks (VANETs) are vital to enhance traffic management and safety in Intelligent Transportation Systems (ITS). However, misbehaving nodes can compromise these benefits by introducing and spreading false information, creating safety risks such as accidents and phantom traffic jams. This paper explores machine learning techniques for detecting misbehavior in VANETs, focusing on speed and position-related attacks. The proposed model demonstrates robust detection capabilities against speed and position offset attacks, utilizing features derived from basic safety messages (BSMs). Evaluated on the VeReMi dataset, optimized tree-based models achieve a detection accuracy of 91%, with a 36% reduction in training time. This dual emphasis on security and environmental sustainability highlights the role of misbehavior detection in creating more safer, efficient, and eco-friendly transportation systems.