Focusing on the research results of psychologists and epistemologists it is an open issue whether people who commit suicide do so as a result of free will or if they suffer from chronic or occasional depression. During the decision making process of a potential suicide the facial features, the voice frequencies and the body movement gestures express the depressive emotional state of a specific individual. In metro stations and other public space there are surveillance systems which can capture facial expressions, body movement and speech recognition of an individual. In this paper it is proposed a design of an information system architecture which can predict whether an individual intends to commit a suicide in an urban metro station or not, in a critical period of time.