Introduction Ulnar tunnel syndrome at the elbow is a common pathology. The ultrasound cross-sectional area is a well-known metric widely accepted in radiology for the description of nerve entrapment. However, the pathological cut-off value remains challenging. The objectives of this study were to (1) describe the ultrasound cross-sectional area measurement of the ulnar nerve at three locations, and (2) to evaluate the inter-observer reliability by two independent ultrasonographers. Methods One-hundred ulnar nerves of 50 asymptomatic individuals were scanned using B-mode and power Doppler ultrasonography. The ultrasound cross-sectional area measurements of the ulnar nerve were performed at three different levels: 2 cm proximal to the epicondyle, at the level of the epicondyle, and 2 cm distal to the epicondyle. Results In our healthy population, we found 21, 24 and 7% of ultrasound cross-sectional area ulnar nerve > 8 mm², respectively, at three different levels of measurement and 4, 7, and 0% US-CSA ulnar nerve > 10 mm². The intraclass correlation coefficient measured at three different site levels were good (0.7943, 0.7509) to moderate (0.5701). Conclusions Almost one-quarter of our healthy population had an ultrasound cross-sectional area ulnar nerve more than 8 mm² and few more than 10 mm². A cut-off of ultrasound cross-sectional area ulnar nerve measurement more than 10 mm² could be considered as pathological. No abnormal elbow ulnar nerve vascularization has been seen. This is the first step towards normal B-mode ulnar nerve values at the elbow to further detect pathological US findings as ulnar nerve entrapment.