To address the challenging problem of unmanned aerial vehicle (UAV) path planning in complex environments, this paper proposes an improved Good-point set Whale Optimization Algorithm (GWOA) to overcome the limitations of the traditional Whale Optimization Algorithm (WOA) , including excessive dependence on the global best solution, insufficient population diversity, and weak exploitation capability in the later search stage. In the proposed algorithm, a good-point set strategy is first introduced during population initialization to enhance solution diversity. Then, a dynamic adaptive weighting mechanism is employed to accelerate convergence while reducing the risk of premature stagnation. Furthermore, an optimal neighborhood perturbation operator is designed to strengthen local exploitation ability, effectively compensating for the deficiencies of WOA in global search accuracy and convergence speed. Experimental studies are conducted in two stages: performance verification using benchmark test functions, followed by comparative simulations in three types of UAV flight path planning scenarios. The results demonstrate that, in the first mountainous scenario, GWOA achieves performance improvements of 21.17%, 7.86%, 5.34%, and 5.83% compared with the Artificial Bee Colony algorithm (ABC) , Grey Wolf Optimizer (GWO) , Whale Optimization Algorithm (WOA) , and Dung Beetle Optimizer (DBO) , respectively. In the second scenario with no-fly zones, the corresponding improvements are 23.56%, 3.82%, 9.36%, and 12.92%. In the third and more complex environment, the performance gains further increase to 35.56%, 18.96%, 4.55%, and 4.80%. Combined with dynamic obstacle avoidance tests, the results indicate that the proposed GWOA not only exhibits superior convergence speed and solution accuracy, but also consistently generates higher-quality paths under diverse constraints, demonstrating strong robustness and environmental adaptability.