A unified framework for the integration of the artificial potential field (APF) theory and the navigation for the search of sensory target sources is presented. This global framework is based on modelling both APF navigation and sensory search as a functional optimization problem, making possible to apply a simple gradient operator that allows a unified navigation. The novel idea of generating steering angles tangent to the potential field curves instead of the conventional use of normal forces is also introduced. Finally, an optimal holonomic kynematics has been implemented in a prototype vehicle. Experimental testing with light and temperature sources and with sonar sensors for obstacle avoidance has shown the efficiency and practical interest of this unified framework.