In the last years there has been an increasing interest in developing robots that are able to perform tasks cooperatively with humans instead of replacing them, for instance a robot which works as a guide in a museum. Until now the developed service guide robots only navigates through the aisles of the museums without concerning the people's movement. Then, the main purpose of the work presented in this article is to develop a nonlinear controller to allow a robot to guide a person through a specific environment until it reaches a desired point taking into consideration the movement capability of the guided person. Besides the controller a person detection and tracking is also necessary in order to accomplish the guidance task. The person detection and tracking are done by using omnidirectional images. The stability of the proposed controller is proved through the Lyapunov analysis. Some experiments are presented to show the effectiveness of the proposed controller.