The PhD work will address complex machining features recognition and similarities measure to reuse the machining process knowledge for complex parts in aeronautics. The approach will be feature-based. First, we will define a new classification of machining features in the context of parts in aeronautics. This classification will consider shape data/information and other relevant information such as PMI or NC machining. A new approach that combines reverse engineering techniques, geometric processing and classical graph-based feature recognition will be investigated. Then, for each class of features we will define a macro process plan that could be refined and would serve as the basis for shape comparison and retrieval. Data mining and deep learning techniques will be investigated.