Background. To develop a simple and cost-effective method for the detection and genotyping of high-risk human papillomaviruses (HPV) using seminested polymerase chain reaction (PCR) and reverse hybridization. Methods. Cervical swabs for HPV testing were collected from 127 women with normal cervical cytology and 57 patients with cervical lesions of various degrees. After DNA isolation, PCR amplification was first carried out using MY11 and MY09/HMB01 primers, then labeled by seminested PCR using the first PCR products and MY11/bioGP6+ primers. One fifth of the second PCR products were resolved by gel electrophoresis. Genotyping for high-risk HPV was done separately, using the remaining products, by a high-risk HPV chip, which contained 13 type-specific oligonucleotides on a nylon membrane. The final result was detected by colorimetric change on the chip under direct visualization. Results. High-risk HPV DNA was detected in 19 (15%) of 127 women with normal cervical smear cytology, in 26 (89.7%) of 29 patients with cervical intraepithelial neoplasia (CIN), and in 27 (96.4%) of 28 patients with invasive cervical carcinoma. Multiple high-risk HPV infections were detected in five cases. HPV type 16 was the most frequent type of infection, comprising 34.5% and 53.6% of the patients with CIN and invasive carcinoma, respectively. The samples without a visible 190-bp band on electrophoresis exclusively showed negative hybridization results. This method could detect one to two copies of the HPV-16 genome derived from one SiHa cell. The overall sensitivity of HPV detection was 25 to 50 copies of HPV genome for each specimen. Thirteen high-risk types and twenty-four different types of HPV DNA showed specific hybridization without any cross-reaction. Conclusions. Our results demonstrated the feasibility and optimistic prospects for this simple and cheap method of high-risk HPV genotyping. This technology can be easily set up in a routine molecular laboratory and would probably be of great value in cervical cancer prevention programs.