Relative coordination has predominant advantage in uncertain environment path planning problem because it denies much of redundant information of the dynamic environment while avoids the rotation transformation which is a big burden in the traditional methods. Based on relative coordination, this paper proposes a trimmed Ant Colony Optimization (ACO) algorithm for three dimensional trajectory generation, i.e., path planning for target-pursuing and obstacle-avoiding (TPOA). We construct the city-map directly on the acceleration of the pursuing vehicle, an aerial robot. Then the robot's kinematics and the dynamics constraints will be embedded conveniently into the ACO formulation. As a result, the searching space in the ACO formulation shrinks greatly and the robot's movement is real and more feasible for application. This novel approach is verified by various simulations and the results show the good performance of efficiency.Index Terms -three dimensional path planning, aerial vehicle, constraints, dynamic environment, ACO.