CNC machine tools are now urgently needed to solve the problem of efficiency improvement and accuracy assurance,and it is difficult for traditional CNC machines to predict and optimize the error accurately and comprehensively under various machining conditions.A new idea and implementation method was proposed to improve the overall performance of CNC machine clusters by applying manufacturing big data,to enable machine clusters to learn and optimize precision at a rate that could not be achieved by a single machine tool,and to study the characteristics and characterization methods of manufacturing process data.The experiments show that the machine clusters mutual learning precision optimization machining method driven by big data model is significantly better than the common traditional CNC machining method in terms of machining accuracy and surface quality,and high efficiency and high quality machining of CNC milling is realized.