The recent development of new offshore projects in pre-salt deepwater fields has placed offshore loading operations as the main production outflow alternative, increasing the operational complexity and risks. Numerous dangerous situations are associated with oil offloading, such as the messenger line transfer during the mooring stage. Nowadays, this critical task is realized by launching a thin messenger cable using the pneumatic line throwing apparatus. This is a complex and slow process since the operation usually occurs with the ship opposite to the wind. This work proposes a hybrid flight methodology based on computer vision and sensor fusion techniques for autonomous unmanned aerial vehicles (UAVs). The UAV takes off from an oil rig and precisely reaches a specific point in the shuttle tanker without using expensive positioning devices and augmenting UAV’s orientation (yaw) precision since the compass can suffer from severe interference due to naval metallic structures near the vehicle. The proposed framework was tested in a realistic simulated environment considering several practical operational constraints. The results demonstrated both the robustness and efficiency of the methodology.