OBJECTIVE: To investigate the possibility for human papillomavirus (HPV) infection to be a predictable signal for the carcinogenesis of oral mucosa by comparing the prevalences of HPV in each stage of oral mucosal carcinogenesis and to compare the sensitivity differences of the two methods in detecting HPV infection in oral cavity. METHODS: The hybrid capture (HC-II) was used to detect infection of HPV in 255 samples taken from 12 cases of healthy oral mucosa, 211 cases of patients with pathological diagnosis and 32 cases of patients with clinical diagnosis. The diagnosed cases included 8 cases of benign lesions of the oral mucosa, precancerous lesions [74 cases of oral leukoplakia (OLK) with hyperplasia and 42 cases of OLK with oral epithelial dysplasia (OED)], 91 cases of precancerous condition [oral lichen planus (OLP)] and 28 cases of oral squamous cell carcinoma (OSCC). And in situ hybridization (ISH) was used to detect infection of HPV in 33 cases of OSCC and 76 cases of OLK, including 30 cases of hyperplasia, 15 cases of mild OED, 15 cases of moderate OED and 16 cases of severe OED. RESULTS: The prevalence of HPV in OLP samples was higher (12.12%, 8/66) than that of OLK (2.59%, 3/116) (χ(2)=4.666, P=0.031) and OSCC(7.14%, 2/28, χ(2)=0.513, P=0.474). The prevalence of HPV in OSCC (7.14%, 2/28) was higher than that of OLK (2.59%, 3/116), and no significant difference was found. There was only one case of smoke spot and statistical analysis was not carried out. ISH was used to detect type 16/18 and type 31/33 HPV DNA in 109 cases of oral mucosal lesions in paraffin sections and only one case of OSCC was HPV positive. Thirty-seven cases were detected by HC-II and ISH methods at the same time. The same negative results by the two methods were found in 94.6% samples (35/37). In the other two samples, one was OSCC with early infiltration and the other was OLK with hyperplasia, The HC-II results were positive while the ISH results were negative. The patients with OLP and HPV testing results were followed up and the average follow-up period was (36.2 ± 10.5) months. It was found that three of them had a malignant transformation, and the malignant transformation rate of HPV positive patients was 12.50% (1/8), which was higher than that of HPV negative patients (3.45%, 2/58), and the difference was not statistically significant, P=0.249. CONCLUSION: HC-II assay was more sensitive in detecting HPV infection of oral mucosal lesions than ISH. The results of this study showed that there was insufficient evidence for taking HPV infection as a predictor of OLK carcinogenesis. Patients suffering from OLP were in a precancerous condition. The prevalence of HPV in OLP patients of this study was higher than that in OLK and OSCC patients, suggesting that for some reason, OLP patients were susceptible to HPV. HPV testing can be considered as routine in patients with OLP, and HC-II assay was recommended. And patients with OLP and HPV positive should be followed up regularly.