Abstract: The existing calibration methods of gamma ray dose (rate) meters have the disadvantages of long distance, long period and low efficiency. To overcome these disadvantages, the Institute of Nuclear Physics and Chemistry of the Chinese Academy of Engineering Physics applies a machine learning prediction method using sample instruments to carry out a research on field calibration technology for portable gamma dose (rate) meters based on minitype reference radiation. A field calibration device for gamma dose (rate) meters for radiation protection was successfully developed, which achieved a satisfactory result with a measurement standard uncertainty of no more than 5%. This paper also presents a new method for determination of the conventional true value of gamma ray air kerma, and provides a novel model equipment for field calibration of the portable gamma dose (rate) meters.