This work proposes a fuzzy approach for path following of a nonholonomic mobile robot, based on the information of a frontal camera. The proposed methodology is divided in three stages. The first stage gets the image of the frontal camera and processes the image to detect and isolate the desired path to follow and eliminate non-useful information. The second stage estimates the orientation for different sections of the path to follow. Finally, in the last stage, a fuzzy system is designed and simulated to control the steering direction of the mobile robot. We show the design, simulations, and experiments using the fuzzy controller. The results are evaluated and discussed in terms of quantitative metrics.