For a visual servoing system with an eye-to-hand configuration, there is often a trade-off between the camera’s field of view (FOV) and visual positioning accuracy. This study presents an innovative visual servoing system that employs dual robotic arms performing distinct functions, coupled with a novel multi-distances observation-based visual servoing technique. To ensure that the target remains within the camera’s FOV, two distinct image path planning strategies are developed for each robotic arm. These strategies are notably independent of system parameters, which enhances their versatility. Numerical simulations and experimental results demonstrate that the proposed method effectively mitigates the conflict between the camera’s FOV and positioning accuracy, making it well-suited for high-precision assembly tasks, while also providing a new approach for enhancing the intelligence of visual servoing systems.