To determine the diagnostic yield of ultrasound-guided core needle biopsy (US-CNB) in cervical lymphadenopathy and identify the factors influencing the diagnostic accuracy of US-CNB. We retrospectively reviewed the records of 6,603 patients with cervical lymphadenopathy who underwent 6695 US-CNB procedures between 2004 and 2017. Adequate specimens were obtained in 92.19 % (6,172/6,695) of cases. Most lymph nodes (67.65 %) were malignant (metastatic carcinoma 4,131; lymphoma 398). The overall accuracy of US-CNB for differentiating benign from malignant lesions was 91.70 % (6,139/6,695). Among biopsies in which adequate material was obtained, the sensitivity, specificity and accuracy of US-CNB were 99.70 %, 100 % and 99.46 %, respectively. The success or failure of US-CNB for the diagnosis of lymphadenopathy was significantly correlated with node size, nature (malignant vs. benign), and location as well as penetration depth, but not with needle size (p = 0.665), number of core tissues obtained (p = 0.324), or history of malignancy (p = 0.060). There were no major procedure-related complications. US-CNB is a safe and effective method of diagnosing cervical lymphadenopathy, and our findings may help optimise the sampling procedure by maximising its diagnostic accuracy and preserving its minimally invasive nature. aEuro cent US-CNB is useful for the diagnosis of cervical lymphadenopathy. aEuro cent US-CNB is safe to perform on lymph nodes located near vital structures. aEuro cent Larger, malignant, level IV lymph nodes yield sufficient tissue samples more easily.