Improper overtaking behavior is one of the main causes of traffic accidents on highways. To address the challenges in designing state machine models for overtaking scenarios, which often suffer from high complexity and insufficient generalization, this paper introduces a dynamic non-cooperative game model to analyze the interactive behaviors between vehicles. Additionally, an overtaking decision-making model is developed by considering different driver styles. Firstly, drivers are classified into three styles — aggressive, normal, and conservative—using factor analysis and $ K$-means clustering. Subsequently, the Stackelberg game theory is employed to describe the interaction between the ego vehicle and the obstacle vehicle. A game cost function incorporating safety, comfort, and traffic efficiency is constructed, and the optimal overtaking decision is derived by integrating different driving styles. Furthermore, a quintic polynomial lane-changing trajectory is studied, taking into account the influence of different driving styles. A decision evaluation function that meets multiple requirements is established to determine the optimal lane-changing time under various combinations of driving styles. Finally, the effectiveness of the proposed decision-making model is validated through co-simulation using PreScan and Simulink across multiple scenarios. The aim is to assist intelligent vehicles in making human-like overtaking decisions, thereby improving traffic efficiency and driving safety.