Abstract: With the rapid development of automotive technology, vehicle driving behavior evaluation has garnered increasing attention from researchers. This paper proposes a driving behavior scoring model based on onboard positioning data. Key parameters including latitude/longitude, speed, and directional angles were comprehensively collected to calculate ten behavioral indicators such as speeding, rapid acceleration, hard braking, and sharp turns. These indicators were converted into percentile scores via hyperbolic tangent functions. An improved analytic hierarchy process quantified indicator weights for comprehensive behavior scoring. Extensive experimental validation demonstrates the model’s high adaptability and accuracy across diverse driving conditions, effectively predicting traffic violation and accident risks. The model provides traffic authorities with real-time monitoring tools and precise behavioral analytics, showing significant practical value in cross-disciplinary applications.