High risk human papillomaviruses (HR-HPVs) are associated with increased risk of normal cervical cells developing to dysplasia and cervical carcinoma. Therefore, HR-HPV DNA testing can predict an endpoint of cervical carcinogenesis that is earlier than the development of cervical abnormalities. Not only the sensitivity of methods but also the amount of HPV DNA are very important and might be parameters to distinguish HPV detection. In this study, we evaluated the effects of primer sets and the polymerase chain reaction (PCR) performance with low viral load samples with normal cervical cytology (140 samples) and mild dysplasia (140 samples) using two consensus primers MY09/MY11 and GP5+/6+. The PCR was performed with single and nested PCR. Positive samples with both primer sets were then HPV genotyped by dot blot hybridization. Results showed higher sensitivity of single PCR using primer GP5+/GP6+ than primer MY09/MY11. HPV DNA was detected in 15% (21 of 140)and 20.7% (29 of 140) of normal cervical samples, respectively. For mild dysplasia samples, HPV DNA was detected in 37.1% (52 of 140) with MY09/MY11 and 50% (70 of 140) using GP5+/GP6+. In normal cervical samples, the positivity rate was increased to 38.5% (54 of 140) by nested PCR using primer GP5+/6+, but only 2 mild dysplasia samples that were negative by single GP5+/6+ were positive by auto-nested PCR. These results suggested that, in low viral load samples, the sensitivity of HPV DNA detection depends not only on primer sets but also PCR performance. HPV 16 was the most common in mild dysplasia samples (20.8%), whereas HPV type 58 was found in 11.1%. This study suggested that nested PCR might be necessary for HPV DNA detection in cervical samples of women participating in cervical cancer screening.