The real-time computer graphics rendering processes embody complex state transitions and fast dynamics amidst observable steady state behavior. This paper presents a systematic framework by which control principles are introduced for tracking control of the rendering processes. The control theoretic approach surpasses existing heuristics and algorithms which are essentially open-loop estimations. Furthermore, both the data-driven methodology and linguistic based control strategy proposed in this paper circumvents technical difficulties posed by rapid advancement in current graphics hardware and software architectures.