This paper proposes a head pose identification method using nose’s relative position information in a face region and develops a game system based on visual head pose identification techniques to control a virtual robot walking in a virtual maze environment. Adaboost classifiers are used to detect a face and give a rough classification. A nose classifier is used to detect noses in a detected face region. The nose’s relative position information is utilized to further determine a specific head pose. A method for selecting the best nose area is also proposed to solve the problem that multi-nose areas may be detected in one face region. Experiment results show that by using these techniques correct head poses can be acquired and applied in the game system.