To address the challenges in UAV(unmanned aerial vehicle) trajectory planning for forest inspection, such as slow convergence and susceptibility to local optima, a trajectory planning method based on the gravitationally reconfigured Keplerian optimization algorithm was proposed. The complex trajectory planning problem was transformed into multiple multidimensional optimization problems by constructing a three-dimensional environment and a trajectory cost model. In the planetary gravity definition of the Keplerian optimization algorithm, the influence of other planets was comprehensively considered to enhance the spatial search diversity. In the exploration phase, a novel position updating strategy incorporating the Levy flight mechanism was introduced to improve the algorithm's ability to escape local optima. In the exploitation phase, a position updating strategy inspired by the particle swarm optimization algorithm was adopted, where the sun's position was regarded as the global optimum, thereby improving convergence. The proposed method was validated in both simple and complex forest environments. Experimental results show that, compared to the particle swarm optimization, whale optimization algorithm, and standard Keplerian optimization algorithm, the GRKOA exhibits superior trajectory planning performance and robustness in complex environments, effectively satisfying the demands of most forest inspection tasks.