Objective: To explore the correlation between characteristics of high-frequency ultrasound features and neuroelectrophysiological parameters of type 2 diabetic peripheral neuropathy (DPN). Methods: The patients with type 2 diabetes mellitus (T2DM) who were treated in the hospital between January 2019 and December 2020 were selected as research subjects, and they were divided into group A (n=43) and group B (n=57) according to whether peripheral neuropathy occurred in patients, and the healthy subjects during the same period were selected as group C (n=40). All participants underwent high-frequency ultrasound examination of the 5 sites of the median nerve of the right upper limb, and the cross-sectional area (CSA) was calculated. Neural-electrophysiological test (NET) was only performed on patients with T2DM, and the correlation between CSA and neuroelectrophysiological parameters [distal motor latency (DML) of wrist-abductor pollicis brevis, motor conduction velocity (MCV) of elbow-wrist, sensory conduction velocity (SCV) of finger 1-wrist, sensory nerve action potential (SNAP) of finger 1-wrist, F wave average latency (F-AL)] were analyzed, and receiver operating characteristic (ROC) curve was used to analyze the diagnostic value of high-frequency ultrasound parameter CSA on DPN at different levels. Results: The CSA values of group A, group B and group C at elbow joint (N1), 4–5 cm above the rasceta (N2), pisiform bone (N3), unciform bone (N4), pronator teres muscle (N5) were all decreased successively (P<0.05). The CSA values at pisiform bone and unciform bone of group A were higher than those of group B (P<0.05). The proportions of blood flow, mesh-like structure disappearance and internal echo reduction were gradually decreased among group A, group B and group C (P<0.05). Compared with group B, DML and F-AL of group A were higher while MCV, SCV and SNAP were lower (P<0.05). CSA was negatively correlated with SNAP, SCV and MCV, and was positively correlated with DML (P<0.05), but had nothing to do with F-AL (P>0.05). ROC curve analysis showed that high-frequency ultrasound had the greatest diagnostic value at unciform bone level, with the best cut-off value of 12.19 mm² and the area under the curve of 0.738. Conclusion: The high-frequency ultrasound parameters and sonograms of patients with diabetic peripheral neuropathy have specific manifestations. The CSA parameters are closely related to the neuroelectrophysiological parameters, and the high-frequency ultrasound CSA shows a higher diagnostic value at the level of the hamate bone, which can be used for early diagnosis of DPN.