<p>Autonomous vehicle path planning often confronts a dichotomy: global planners ensure optimality but lack dynamic adaptability, while local planners excel at real-time obstacle avoidance but are prone to local minima. To address this challenge, this paper proposes a Hierarchical Adaptive Path Planning Framework (HAPPF) that synergistically integrates an enhanced A* algorithm with an enhanced Dynamic Window Approach (DWA). The A* algorithm is optimized through multiple strategies, including the introduction of obstacle buffers, turning penalties, an adaptive heuristic function, and an improved novel 5-neighbor search strategy. Concurrently, the DWA’s evaluation function is augmented to include path smoothness and time costs, and is supplemented by a core RITR-DE strategy for sophisticated dynamic obstacle handling. Within this framework, the enhanced A* generates a globally optimal path to guide the DWA in performing local navigation and real-time obstacle avoidance. Extensive comparative and simulation experiments validate the framework’s superiority. The proposed HAPPF comprehensively outperforms the traditional A*-DWA framework in a complex 50 × 50 simulation. In terms of performance, it reduced average planning time by 61.1% and the number of turning points by 68%. In terms of reliability, it achieved a perfect record of 100% planning success and zero collisions in 50 repeated physical tests.</p>