An automatic diagnosing method for the autonomic nervous system and a device thereof are described. The device comprises a sensor, a computing device and an output device. An electrode of the sensor is adhered to the skin surface of a subject to detect and output the heart beat signal of the subject. The computing device collects the heart beat signal from the electrodes through the signal collection leads. The computing device further amplifies, filters, digitizes and transforms the heart beat signals into a plurality of heart rate variability parameters. Further, calculations, validation and analysis are performed on these parameters. After selecting a corresponding diagnosis description statement from a lookup table in a database, an examination report, which incorporates the diagnosis description statement and the heart rate variability parameters is output.