Filter debris in a lubricating system contains important wear information of an aero-engine, and its quantitative study through image analysis is an effective means for engine wear condition monitoring and fault diagnosis. In order to meet the practical requirements of quantitative filter debris analysis for a certain type of aero-engine, a special automatic device for oil filter rinsing and debris collecting is designed. The collected debris is distributed on a fabric film according to its magnetic properties, and the digital image of the debris is obtained through a charge coupled device(CCD). The features of the debris image are extracted through the 2D maximum entropy genetic algorithm image thresholding, and then these features are used to obtain wear degree, which reflects the wear condition of an aero-engine. In the machine learning process, by means of analyzing a lot of case images of filter debris, the relationship between features of filter debris images and wear degrees of aero-engines is established through multiple regression. In the test process, the features of the filter debris image taken from the CCD are obtained via image analysis, and then the wear degree of an aero-engine is calculated directly through the regression formula. The wear degree derived from the filter debris image analysis can be used to diagnose the wear condition of an aero-engine by comparing it with the threshold of wear degree. Finally, actual application of the quantitative filter debris analysis during the maintenance of an aero-engine certifies the effectiveness of this new technique.