STEP-NC is a smart standard, developed by the International Standard of Organization ISO, to substitute the ISO 6983 G-code, because, the language G-Code, normally used for Computer Numerical Control (CNC), is qualified to be not able to links CAD/CAM/CNC digital chain and ensure the exigencies of modern intelligent manufacturing in terms of tractability, interoperability, flexibility, adaptability, and extensibility. Therefore, the first objective of this paper is to design and implement a Computer Automatic Aided Process for Turning process, designated by CAPP-Turn, to ensure machining of rotational parts within this modern vision. However, to achieve CAPP-Turn system, it is compulsory to build a robust Automatic Manufacturing Features Recognition AMFR module to establish a full communication between the first two links of the digital chain which are Design CAD and Manufacturing CAM. that's why, by using a hybrid graph-rules method, the second objective of this works is focused on elaboration of a new consistent-fast algorithm that allow extraction of the machining turning entities for parts with most efficiency and complex geometry. In fact, in the literature, most of the presented AMFR systems are restricted of external turning process and cannot handle parts with complex geometry and interacting features. Moreover, the frontal turning features are almost neglected in most of these systems, despite their importance for fulfilling certain functions in mechanical systems. This article, in first, details the global architecture of the CAPP-turn and describe clearly trades between the CAD part and STEP-NC output file. In second, it explains model of the Automatic Manufacturing Feature Recognition (AMFR) system. This system encompasses: (i) a parser module that translates geometric and topological data, from STEP AP203 CAD file, into Python entity class’s objects; (ii) an AMFR that analyses the created-objects and applies predefined-rules to construct all possible turning machining (iii) a Module capable to select external features from internal, frontal features from axial and handle interacting features from the simples. Afterwards, these steps, the AMFR gives all suitable sequencings for part machining. At the end, with a goal to demonstrate the potential advantages and power of proposed the proposed AMFR, a selective part is chosen for the test. The result shows that AMFR performs well on recognizing all types of features indifferently of their types: Internal or external, axial or frontal, simple or interacting.