Fine-needle aspiration (FNA) biopsy of the parapharyngeal space (PPS) is a diagnostic challenge and sampling is often done without image guidance, often trans-orally. Primary PPS tumors are rare, and there is a broad differential diagnosis. The accuracy of PPS FNA, in particular without CT-guidance and using liquid-based cytology (LBC), has not been well studied. Pathology records from our institution (a 1,100 bed Canadian academic tertiary care centre) were searched to identify all patients who underwent PPS FNA from September 1991 to August 2009. The FNA diagnosis was compared to the gold standard of subsequent histopathology or long-term clinical follow-up. Of 36 FNAs, 3 employed image guidance. Eleven (31%) FNAs were nondiagnostic. In the 25 diagnostic FNAs, there was sensitivity 89%, specificity 94%, PPV 89%, NPV 94%, and accuracy 92% for the diagnosis of positive or negative for malignancy. A correct specific diagnosis was made in 9/25 (36%) cases. The nondiagnostic rate was significantly higher (P < 0.025) in FNAs prepared as conventional smears (9/17 = 53%) versus LBC (2/19 = 11%). A specific diagnosis was made significantly more often (P < 0.05) with LBC (8/19 = 43%) versus conventional smear (1/17 = 5.9%). One minor complication from FNA occurred. In conclusion, PPS FNA is safe and accurate for the diagnosis of malignancy. The rate of reporting a specific diagnosis is low. Nondiagnostic FNAs are common. There are more specific diagnoses and fewer nondiagnostic tests with LBC than with conventional smears. Improved specimen quality with LBC is likely a factor.