A high-visibility path planning method is proposed for tracking moving targets in complex and dense environments. First, in order to respond to target motion changes in advance, an A* path search method based on target state prediction is designed. The target distance and obstacle distance are introduced as a heuristic cost to generate a path that avoids obstruction by obstacles and obtains the initial trajectory. Then, to address the problem that targets are easily occluded in complex and dense environments, target visibility constraints and dynamic safe distance constraints are introduced into trajectory optimization. According to the distribution of the target and obstacles in space, the trajectory is adjusted to reduce the probability that the target is occluded by obstacles. Combined with dynamic constraints and collision constraints, MINCO is used to complete trajectory optimization and generate a flight trajectory with minimal control effort. Final, simulation experiments prove that the algorithm proposed has a target visibility rate of 96.83%, which is higher than Elastic-Tracker and Fast-Tracker. Real-world experiments demonstrate the engineering feasibility and robustness of the proposed algorithm.