Purpose : This study describes the development of a real-time video-based eye movement measurement device and evaluated its performance using repeated measurements with a printed eye model. The aim was to examine the stability and repeatability of pupil center detection using low-cost hardware and open-source software. Methods : Infrared-illuminated grayscale eye images were captured with a high-speed camera, and a real-time algorithm detected the pupil and extracted its center coordinates. A printed eye target was placed at a fixed distance, and ten repeated trials (900 frames each) were recorded. The pupil detection success rate and the standard deviation (SD) of the center coordinates were calculated. Results : The overall detection success rate was 97.3%. The SD of the pupil center coordinates remained below 1 pixel (0.46±0.05 pixels horizontally and 0.52±0.04 pixels vertically). No systematic drift was observed in the time-series distribution of coordinates. Conclusion : The device showed high repeatability and reliable pupil detection in a printed eye model. The low-cost system and adjustable algorithm offer a flexible platform for early-stage development without IRB approval and may be expanded to analyze pupil diameter, vergence, and pupillary oscillation in future work.