Background The role and extent of neck dissection in patients with parotid metastatic cutaneous head and neck melanoma remain unclear. The aims of this study were to determine the incidence and patterns of cervical node involvement in patients with parotid metastatic melanoma, and to determine if a limited lymphadenectomy of the clinically negative neck is appropriate. Methods Patients who underwent parotidectomy and neck dissection for clinically apparent parotid metastatic melanoma, irrespective of neck status, were identified from two prospectively maintained databases. Results A total of 276 patients fulfilled the study criteria. Median follow‐up was 23 months. A total of 185 necks were clinically negative, 82 were clinically positive. A total of 36 elective neck‐dissection specimens harbored occult metastases; these were found in levels I (16.7%), II (58.3%), III (36.1%), IV (13.9%), and V (30.6%). Regional recurrence occurred in 32 patients with a clinically negative neck, the majority being in‐transit metastases (n = 15). Only one case of recurrence could have potentially been avoided by a comprehensive lymphadenectomy. Conclusions In patients with clinically apparent parotid melanoma metastases, elective comprehensive neck dissection reduces failure rates in cervical nodes, and provides more accurate staging and prognostic information. However, our findings support the emerging trend for more limited elective neck dissection. Levels I and IV can probably be safely omitted.