This study aims to suggest a methodology which combines Magnetic Flux Leakage (MFL) and Convolutional Neural Networks (CNNs) technologies in order to quantify internal and surface defects in wire ropes. Wire ropes are widely used as a key component in most engineering structures, which makes it necessary to pay attention to their condition as the structural integrity of these components is critical when it comes to safety and performance issues. In fact, classical NDT techniques such as visual and ultrasonic ones often fail when it comes to locating internal defects or providing an accurate quantitative assessment of those already detected. MFL allows detecting changes in magnetic flux caused by defects, but, at the same time, does not allow making a distinction between internal and external defects and cannot accurately measure parameters of defects (e.g., their width, depth, and cross-sectional loss). Thus, MFL is combined with CNNs to provide more efficient and automated inspections. A CNN algorithm, trained on MFL data, should be able to recognize features related to certain defects and provide a precise estimation of defect parameters.The effectiveness of the system is validated through both simulation and experimental data, demonstrating its potential to enhance safety in industries that rely on wire ropes.