We propose a system that performs automatic laser spot detection and tracking. Our laser spot detection approach is based on the HSV color model, thresholding, and center of mass. For tracking, a system of quadrants was used to move a camera depending on the center of mass. Also, the distance to the laser spot is calculated using a stereo vision configuration and geometric relationships. The proposed architecture is composed of an embedded computer that controls a low cost, semi-autonomous sentry robot attached to it. Our results show a robust real-time system that can detect a laser spot and track it, with a distance error lower than 0.80% on average and an execution time of 62 milliseconds per frame, that can be used in different applications.