Herpes simplex virus infections may be diagnosed by several techniques, including conventional cell culture and the polymerase chain reaction (PCR). This prospective study compares the analytical performances and usefulness of an in-house real-time PCR method and the Light Cycler HSV (1/2) detection kit (Roche Diagnostics, Mannheim, Germany). The results of both PCRs were then compared to those obtained by conventional cell culture. A total of 313 samples were tested (70 dermal samples, 81 cerebrospinal fluids (CSF), 47 ocular, 42 anogenital, 34 throat swabs, and 33 oral samples, 3 whole blood, 2 biopsies, and 1 bronchoalveolar lavage). Samples for molecular assays were extracted twice with the MagNa Pure instrument (Roche Molecular Biochemicals, Mannheim, Germany) and tested blind in parallel by the two PCR methods. Most (226) samples were also examined by cell culture. Forty three samples were found positive by both PCRs, whereas 267 were negative. The HSV-1 and -2 typing of positive samples was identical. Three of the samples were positive in the in-house PCR and negative in the Light Cycler HSV (1/2) detection kit. There was no statistically significant difference between the two tests. Only one sample gave an invalid result due to negative PCR and negative internal control result. Seven samples were positive by both real-time PCRs and negative in conventional culture. The PCRs were significantly (P < 0.05) more sensitive. The results show good agreement between the two real-time PCR methods, with the molecular tests being more sensitive than cell culture.