Mechatronic systems are widely used in many industrial applications for their capacity to improve the production performances. However, due to the complexity of the mechatronic systems, several faults that disrupt the system operating, may occur and specially, the dynamic defects (e.g., electrical, mechanical, design defects • • • ) and the sensors faults (e.g., bias or gain defects). In this paper, classification of some diagnosis methods is presented and a new diagnosis technique, called sparse recovery, is presented. Academic example is developed in order to illustrate the good performance of the proposed method.