Abstract: Due to the presence of floating-base disturbances during aerial grasping by a dual-arm aerial robot, current servoing methods remain inadequate for achieving precise end-effector control. For this problem, a position-based visual servoing (PBVS) scheme which incorporates a self-balancing motion control strategy of dual-arm manipulators is proposed to improve the grasping accuracy of the end-effector. The dual-arm manipulator swings within the x-z plane, so the dynamics model of the aerial robot system is established in this plane. An improved lightweight YOLOv5s-based target detection algorithm is applied to providing the accurate position information of the target by a binocular camera. Considering the force and torque effects on the hexacopter caused by the displacement of the system center-of-gravity (CoG) due to dual-arm manipulator motion, a self-balancing motion control strategy is proposed to minimize the impact of the movement of the dual-arm manipulator during the servoing grasping process, ensuring the stability of the position and orientation of the hexacopter. Consequently, the pose of the end-effector of the manipulator relative to the target object is adjusted through the PBVS strategy, achieving high-precision servoing control. In order to perform the grasping task, the active manipulator moves to the target position, and the passive manipulator moves to the corresponding position subsequently to grasp the target object. The robot operating system (ROS) simulation and physical experiments verify the effectiveness of the proposed PBVS strategy. The experimental results show that the proposed strategy can effectively improve the grasping accuracy of the end-effector of the manipulator.