Type 2 diabetic mellitus (T2DM) patients with sub-clinical cardiac autonomic neuropathy (CAN), which precedes clinically apparent CAN, may have reduced heart rate variability (HRV); yet may remain undetected by conventional autonomic test. The objective of this study was to explore the short-term HRV by power spectral analysis as a tool to detect sub-clinical diabetic cardiac autonomic neuropathy (CAN). This cross -sectional study included 30 recently diagnosed and 54 long-term T2DM male patients (aged 45 to 55 years and without any symptoms of CAN) and 30 age and body mass index (BMI) matched healthy male subjects as controls. This study was done from January 2015 to January 2016 in Endocrinology OPD, BSMMU, who were previously or then diagnosed by the physician at this very OPD following the diagnostic criteria set by the American Diabetes Association Ewing battery of cardiovascular reflex tests (CARTs) was performed to rule out signs of autonomic neuropathy in enrolled patients. Power spectral analysis of frequency domain signals of short term HRV were done by digital data acquisition device RMS Polyrite-D. For statistical analysis, One way ANOVA and independent sample 't' test were used. LF power, HF power and HF normalized units (nu) were significantly lower, and LF nu and LF/HF were significantly higher in long-term T2DM compared to recently diagnosed T2DM (p<0.05) and control (p<0.001). LF power (p<0.01) and HF power (p<0.001) were significantly lower in recently diagnosed T2DM compared to control. Results from this study demonstrate the presence of autonomic dysfunction in CAN asymptomatic long-term T2DM patients. Short-term HRV, thus, seems to be an invaluable tool to detect sub-clinical diabetic CAN.