Abstract: With the advancement of unmanned autonomous system technologies, path planning for truck-drone collaboration still faces challenges such as obstacle handling and kinematic feasibility in complex environments, necessitating in-depth research to improve mission success rates and execution efficiency. An allocation-search-optimization (ASO) algorithm is proposed for task allocation and path planning of truck-drone systems, addressing obstacle factors and agent kinematic feasibility through obstacle avoidance processing and trajectory optimization of agents, respectively. In the initial path search phase, a truck-drone ECBS (enhanced conflict-based search) algorithm is designed to achieve obstacle avoidance among different types of agents in truck-drone systems, named TD-ECBS. In the back-end trajectory optimization phase, a model predictive control (MPC) algorithm suitable for truck-drone systems is constructed to optimize the trajectories of agents within these systems. Obstacle environments are introduced based on existing instance sets, and simulations are conducted in the ROS (Robot Operating System). Results indicate that compared to the algorithms proposed by Boccia and Gonzalez-R, the proposed algorithm increases the critical obstacle density threshold by more than 20% when the planning success rate is below 50% in complex environments with numerous obstacles, while considering agent kinematic feasibility. Even when obstacles and agent kinematic feasibility are not considered, the proposed algorithm reduces the planning time by 6.87% and 4.23% compared to Boccia's algorithm at agent quantity levels of 10 and 20 respectively, demonstrating the efficiency and effectiveness of the proposed algorithm when applied to truck-drone systems in complex unstructured environments.