The consequences of frequent occurrence of turnout fault on the railway, are often very serious, which requires a more complete operation of the state monitoring mechanism and intelligent fault diagnosis methods to improve the safety and reliability of turnout equipment. First, according to the curves of various turnout faults obtained on site, the reference curves of several typical faults were selected as a template curve of diagnosis. Then, according to the similarity based on the Fréchet distance，the similarity degree between the test curve and the template curve was calculated, and the failure mode corresponding to the template curve with the largest similarity was output as the diagnosis result of the test curve. The experiments show that this method does not need a lot of training samples, and it is not necessary to extract the segmentation feature of the action current curve. The diagnostic method has a high accuracy, a high speed, and a strong adaptability.