In this article, an improvement in a system for position measurement of a motorcycle is presented. Position measurement of a mo- torcycle when running presents a difficult problem because loading of measurement equipment would cause changes in the vehicle’s mass and its moment of inertia. Therefore, this paper proposes a novel measurement method that uses omnidirectional cameras to acquire angles relative to fixed camera positions. The method is based on a general stereoscopic positioning approach. The results of previous research have shown that a simple measurement method using image processing techniques could be ap- plied to the position measurement of a motorcycle when running on a figure-of-eight-shaped course around two omnidirectional cameras. The main weakness of this method using two omnidirectional cameras is the large error that occurs near the camera baseline (i.e., the line connecting the two cameras), particularly in the baseline direction. To improve position measurement preci- sion, the author has added two more omnidirectional cameras to the system. Running tests of the proposed system using a real mo- torcycle were executed on a paved area. The experimental results showed that the proposed measurement method is sufficiently accurate to allow it to check the locus of a motorcycle running on a figure-of-eight-shaped course.