Stereo vision–based 3D reconstruction for moving objects has been applied in diverse domains, such as robotic navigation and autonomous driving. However, conventional vision systems often fail to achieve both high reconstruction accuracy for moving objects over a large field of view (FOV) due to imaging degradation and the tendency of objects to exceed the FOV. To overcome these limitations, we propose a dual-galvanometer-based active stereo vision system for large-FOV high-precision 3D reconstruction of moving objects. Firstly, we design a dual-galvanometer-based vision system that rapidly and stably switches camera viewpoints. Following that, we propose a synchronous detection algorithm that can actively capture stereo image pairs and detect moving objects across the enlarged FOV. Then, we build a system-level high-precision 3D stereo reconstruction framework that tightly couples viewpoint switching, synchronized detection, and high-precision 3D reconstruction to improve the reconstruction accuracy and robustness of fast-moving objects. Finally, extensive experiments demonstrate the effectiveness of the proposed system: compared with the fixed-lens FOV, the observable FOV is expanded by 53.7×, while the best-case reconstruction error for a 20 mm moving ball remains below 0.03 mm in the large-FOV workspace.