Introduction: A mass in the salivary gland region often presents a diagnostic challenge with regard to its site of origin (salivary versus nonsalivary), benign or malignant nature, and tissue-specific diagnosis. The present study describes the utility of fine-needle aspiration (FNA) cytology in the diagnosis of these lesions. Subjects and Methods: Over a 6-year period (January 1994 to December 1999), 712 patients aged between 6 months and 91 years (median, 37 years) were subjected to FNA of swellings in their salivary gland regions. Male:female ratio was 1.28:1. The swellings were mostly located in the parotid (323 cases), submandibular (343 cases), and upper cervical region (27 cases). Swellings of oral (5 cases) and sublingual (2 cases) sites were rare. The lesions diagnosed by FNA cytology were compared among the major salivary glands. Cytologic diagnoses were correlated with histology in 45 cases. Results: Benign nonneoplastic lesions were the most common (73%), followed by neoplasms (20%), and those with atypical cytology (1%). Cytologic material was inadequate in 6% cases. Parotid gland region was involved more frequently by neoplasms (27.1%) than the submandibular gland region (13.7%, p < 0.0001). Inflammatory processes affected the submandibular gland region more commonly (42.0%) than the parotid (32.6%, p = 0.0164). Pleomorphic adenoma was the most common neoplasm (61.5%), followed by Warthin’s tumor (12.6%). Malignancies accounted for 10.5% of neoplasms. Frequency of involvement of parotid by Warthin’s tumor (16.7%) was significantly higher than that of submandibular gland (2.3%, p = 0.0191). However, the submandibular gland was more commonly affected by malignancy than the parotid gland (p = 0.0003). Sensitivity, specificity, and diagnostic accuracy of FNA cytology for all neoplastic lesions of the salivary gland were 94.6, 75.0, and 91.1%, respectively. The corresponding figures for malignancies were 60.0, 95.0, and 91.1%, respectively. Conclusion: FNA cytology is very useful for the diagnosis of salivary gland lesions. However, sampling and interpretation errors may occur. The low specificity for the diagnosis of neoplasms as a whole and the poor sensitivity for malignancies found in our study can be attributed to the relatively small number of benign nonneoplastic and malignancy cases with available histopathologic diagnoses.