In the research of non-invasive cardiac coronary imaging with multi-detector CT (MDCT), it is necessary to use a standard cardiac phantom to perform the mass repetitive experiments. The standard phantom can help to investigate the regularity of a technology, and hence provide the instructions to clinic. In this study, an objective evaluation method of coronary stenosis rate, which can process CT images automatically, is developed with a cardiac phantom, to enhance working efficiency and to reduce subjective error. Several experiments were designed to verify the accuracy, the sensitivity and the consistency of this method. Based on the results, we found the accuracy of this objective method is within the clinical acceptable range. This method is sensitive to image quality change, and the results of the method are consistent without being affected by the environment and the operators. This proposed method can help to improve the efficiency of processing mass experimental data based on the cardiac phantom and establish a good foundation for the research on the cardiac coronary imaging with MDCT. This method is also applied to clinical images for tests and acceptable results are achieved in stenosis of low density.