<p>Existing trajectory tracking control methods for mobile robots primarily focus on minimizing pose errors, but they often neglect the impact of path curvature on tracking performance. This paper addresses this gap by proposing a novel trajectory tracking control method based on a Path Evaluation Model (PEM). The first step involves establishing a kinematic model for the mobile robot, followed by the development of the PEM, which integrates both "road condition" (path complexity) and "vehicle condition" (pose error). The path evaluation function is then designed by combining these two factors, with path complexity quantified in terms of curvature and its rate of change. To validate the effectiveness of PEM, we introduce the Path Evaluation Model Controller (PEMC), an adaptive trajectory tracking controller that combines the BP neural network with the Backstepping method. The stability of PEMC is rigorously proven. Furthermore, we provide a detailed analysis of the core parameter ranges in the PEM through simulation experiments, demonstrating the method's robustness. Experimental results show that PEMC improves tracking accuracy by more than 50% compared to traditional adaptive Backstepping controllers, highlighting its superior performance in handling complex trajectories. Additionally, PEMC also reduces computation time, further demonstrating its efficiency and suitability for real-time trajectory tracking in mobile robots.</p>