This paper reviews a theory of robot motion planning and elaborates an application in which neural networks are used to take decisions regarding the orientation of a robot in its search for a destination target.The Simulink is used as a tool to implement the concepts. The neural networks allow obtaining a model of inverse dynamic, indeed the mathematical model of robot. The test was realized with different types of trajectories to validate the methodology proposed. A servosystem was adjusted optimally by the implementationof an heuristic technic.