The aim of this study was to compare and analyze Ewing test, heart rate variability (HRV), and heart rate turbulence (HRT) in the diagnosis of cardiac autonomic neuropathy (CAN) in diabetic patients.Ninety diabetic patients (age 18-78) and 20 nondiabetic control subjects were studied. Standard Ewing test and 24-hour Holter was performed in all participants to evaluate CAN. Patients with Ewing score >/=2 were classified as CAN+.The rate of CAN+ in diabetic patients [44.4% (40/90)] was higher than that in the controls [5% (1/20)] (P < .05). Using the HRV analysis and HRT analysis, rates of CAN+ in diabetic patients were 56.67% (51/90) and 52.22% (47/90), respectively. SD of all normal-to-normal (NN) intervals (SDNN), SD of the average NN intervals calculated over 5-minute periods of the entire recording (SDANN), low frequency power (LF), and turbulence slope (TS) were significantly correlated negatively with Ewing scores. TS (r = -0.68, P < .05) and SDNN (r = -0.58, P < .05) had the strongest correlation with Ewing scores among relevant factors. Combining TS with SDNN as diagnostic criteria for CAN, the diagnostic sensitivity can be increased to 98%.Parameters used for evaluating parasympathetic functions in Ewing test, HR variability, and HR turbulence were found to significantly decrease in CAN+ group. The combination of SDNN and TS showed greater diagnostic value than Ewing test, HRV analysis, or HRT analysis alone.