Part quality is an important aspect in the aerospace industry in order to avoid unpredictable and irreversible damages, and to ensure a long lifecycle. To ensure quality, non destructive testing (NDT) techniques are applied on parts and one of the most frequently used in the aerospace industry is eddy current testing (ECT). However, the latter is still mainly performed manually, enticing high costs in addition to be time-consuming. In this case, reliability and repeatability of inspection results are also limited due to their high dependence on the human operator. As part of an effort to robotize ECT with a 6 DOF manipulator arm and thus, free oneself from the many drawbacks of manual inspections, the authors previously proposed a coverage path planning methodology dedicated to complex aeronautical surfaces. However, orientation of the probe along these paths were not initially considered during this previous work while it is as much important as position information given that the probe must keep a normal orientation with respect to the surface to inspect it reliably. Following on from this previous work, the computation of these orientations along every point of the path is presented in this paper. Furthermore, another coverage path planning methodology is developed to inspect with ECT around a probable defect whose position is assumed to be known a priori. Once positions and orientations along inspection paths are computed, a simulation with a robotic simulation software (MotoSim) is made to generate the joint trajectories of a manipulator arm MotoMan SV3XL and teach it the automated inspection task. Afterwards, experiments using this manipulator with a mock-up representing a probe is realized. After these first experiments, it is observed that the motion of the robot is what one can expect during an efficient inspection around a known indication. et orientations calculées le long de ces chemins, les trajectoires d'un bras manipulateur MotoMan SV3XL qui permettront de suivre ces chemins sont ensuite déduites à l'aide d'un logiciel de simulation robotique (MotoSim). Enfin, un test expérimental sur le robot précédemment simulé est effectué en remplaçant la sonde par une maquette représentative de celle-ci. À la suite de ce test, les trajectoires du robot sont, à première vue, celles attendues pour réaliser une inspection efficace autour d'une indication.