The combination of machine vision and open CNC system contributes to develop various functions of CNC system such as visual inspection, condition diagnosis and machining error compensation. Focusing on the two commonly used software development platforms (Visual Studio, Qt) and three machine vision libraries (OpenCV, Halcon and EmguCV), this paper studies the operating efficiency of software platform for development of open CNC system with machine vision function. The operating efficiency of various schemes is experimentally compared. Finally, the machine-vision-based error compensation efficiency is studied by taking the curve grinding CNC system as an example. The optimization scheme based on operating efficiency is presented, which provides a basis for the selection of software platform for machine vision oriented CNC software.