Parotid lymph node metastasis was supposed to be associated with widespread metastatic disease in the neck. This study was designed to detect the high-risk factors of parotid lymph node metastasis in NPC. Materials/Methods: We retrospectively reviewed the archives of 1096 patients with NPC who underwent IMRT in our hospital from January 2005 to December 2012. 20 patients showed parotid lymph node involvement, in which 3 patients showed bilateral parotid lymph node metastasis (PLNM). Of the 20 PLNM patients, 19/20 patients were N2-3 disease. In addition, one patient with N1 disease was removed. We conducted a 1:2 case-control study. Cases were 22 sides with ipsilateral PLNM (3 pts with bilateral PLNM). We randomly selected 44 patients as controls from N2-3 NPC disease. Potential factors were compared between the two groups, including involved node level, sum of the largest diameter for multiple lymph nodes (SLD) in each level, extra-nodal neoplastic spread (ENS), and node necrosis. In addition, we regarded the level I, Vc (the lateral supraclavicular nodes), and VI as the rare neck areas. Analyses were performed using the statistical analysis software. Results: 20/1096 (1.82%) NPC patients were found parotid lymph node involvement. There was no significant difference between the two groups in age, sex, histology, T stage, and chief complaint. Features of lymph nodes in retropharyngeal area (RPN), level II, III, IV, Va/b, and rare neck area (I, Vc) were analyzed, including involvement, SLD, ENS, necrosis. SLD in level II was larger in case group than that in control group (6.0cm vs. 3.6cm, P Z 0.003). Level II lymph node necrosis, level Va/b and rare neck area involvement were more common in case group (P Z 0.003, P Z 0.034, and P<0.001, respectively). RPN, level III, level IV showed no significant difference in two groups. Receiver operating characteristic (ROC) cure showed that SLD !5cm in level II has a moderate accuracy to select parotid lymph node metastasis with sensitivity of 72.7% and specificity of 72.7% (Az Z 0.73, P Z 0.003). When SLD !5cm in level II, necrosis in level II, involvement in level Va/b, and involvement in rare neck area were put into binary logistic regression analysis, only SLD !5cm in level II (OR Z 4.1, P Z 0.030) and rare neck area involvement (OR Z 3.95, P Z 0.045) were associated with PLNM in NPC patients. Conclusion: PLNM was an uncommon event in NPC patients. SLD in level II !5cm and rare neck area involvement (I, Vc) may be the potentially high-risk factors for parotid lymph node involvement in NPC patients. In addition, IMRT whether to spare parotid should be cautiously reevaluated in NPC patients with high-risk factors of PLNM.