AIMS: To investigate whether tests for cardiovascular autonomic neuropathy (CAN) and 24-hour heart rate variability (HRV) could improve the prediction for outcomes in type 2 diabetes. METHODS: 541 type 2 diabetic individuals performed tests of CAN. A subsample (313) had 24-hour HRV (the standard deviation of all normal RR intervals [SDNN] and the standard deviation of the averaged normal RR intervals for all 5 min segments [SDANN]). Multivariate Cox regressions examined the associations between CAN/low HRV with cardiovascular events (CVEs) and all-cause mortality. The improvement in risk discrimination of adding CAN/HRV was tested by C-statistics and by the Integrated Discrimination Improvement (IDI) index. RESULTS: 25% had CAN, and 17-18% had low HRV, respectively by SDANN-SDNN. Over a median follow-up of 12 years, there were 177 CVEs and 236 all-cause deaths in the whole cohort, and 96 CVEs and 129 all-cause deaths in the subsample. CAN was associated with 40% excess risks of CVEs/all-cause mortality, low HRV was associated with 2-fold higher risks of outcomes. HRV improved risk discrimination for CVEs/mortality with increases in C-statistics up to 0.039 and IDIs up to 25%. CONCLUSIONS: Low HRV was a better predictor of outcomes than tests of CAN, and it improved risk discrimination.