The selection of appropriate cutting parameters is a fundamental factor influencing the machining quality of advanced materials, particularly in applications involving intricate component geometries. Commonly, machining condition control depends on experienced machinists to transfer knowledge into the NC code and processing to the CNC machine controller. The principal cutting conditions consist of cutting speed, feed rate, and depth of cut, which affect the machine power and the quality of machining surface roughness and dimensional accuracy. In an advanced 5 axes CNC machining process, there is limited information for creating the machining parameters under the design condition requirements. The new machine is used for experiments and captures knowledge from time to time. However, the process is expensive and time consuming. This paper proposes a new concept of machining data and knowledge generation, employing empirical data to modify the inference reasoning system of the ANFIS. Instead of adapting knowledge from experts or documents, this paper presents capturing knowledge from experiments and adjusting actual information inside the inference fuzzy rule based system. The contribution is developing the 5 axis milling parameter prediction on the practical knowledge-based system.