To comparatively analyze the methodological efficacy of the polymerase chain reaction (PCR) assay for herpes simplex virus 1 (HSV) detection in tears. This retrospective study reviewed the medical records of 115 patients who were clinically diagnosed with herpes keratitis, and their tear samples were collected for HSV detection. PCR positive rates were analyzed for their dependence on the PCR primers used (conventional PCR primer vs. nested PCR primer), the tear collecting method used (micropipetting vs. collection with schirmer strip), the disease manifestation and the patient's previous medication history. HSV DNA was detected in 23 out of 115 (20%) tear samples. The PCR positive rate in tear samples did not differ depending on the PCR primer or tear collection method used. Typical epithelial lesions showed a higher positive rate (31.4%) than atypical epithelial lesions (10.9%). The previous history of the antiviral agent seemed to affect the PCR positive rate. Although the PCR positive rate was not dependent on the tear collection method or primers, HSV detection in tears using PCR was shown to be a supplementary diagnostic test in typical and atypical herpes epithelitis.