There are several conventional methods for extracting facial feature points. However, these methods are not so good at extracting feature points from face images that were taken with varied poses. In this paper, we have introduced the detection method of the nose position based on the artificial neural networks (ANN). Because of the temperature distribution in nose region is necessary for checking the fatigue, the ANN methods achieved higher detection rate by the suppression learning and the action of rotated face images at learning process. Therefore, we have described that the proposed method is effective for the fatigue measurement.