This paper explains the process of design and implementation of a face detection and tracking system able to locate the face of autism children with low functioning through a depth camera, that establishing the distance and angle relative to a reference point, controls in real-time the movement of the head of a robotic model. The system was developed using the Microsoft Kinect® SDK library, Artificial Vision techniques and the development environment LabVIEW, and makes part of a robotic prototype oriented to therapeutic support and diagnostic of autism spectrum disorder. The results show a good system performance, so the next step will be validate its functioning with real patients and then integrate it on the final model.