To facilitate dynamic deformation measurement of multi-camera video extensometer images, this study proposes a tracking matching method for large-field-of-view multi-camera video extensometer images. The method, grounded in digital image correlation and object point reprojection theory, is applicable to material strain measurement under large deformations. Firstly, the global calibration of multiple cameras was performed, and the matrix transformation relationship between the camera coordinate system and the world coordinate system was solved for each camera. Secondly, the image points of the sequence images acquired by the video extensometer were traversed and tracked between the camera images of the video extensometer by applying the object point reprojection theory. Finally, the equation of multi-camera video extensometer strain measurement was derived, and a multi-camera video extensometer strain measurement platform was built. The measurement results were compared with the conventional strain gauge strain measurement method through tensile experiments on two specimens of Q235 mild steel and aluminum alloy. The results show that the method accurately completes dynamic tracking measurement of image points in multi-camera video extensometer images, with a camera calibration reprojection error of less than 0.05 pixel and average strain errors of 0.170 9% and 0.049 9% for the two specimens, respectively. The results show that the dynamic tracking measurement method of multi-camera video elicitor is basically suitable for strain measurement scenarios with large strain.