Human papillomavirus (HPV) infection is etiologically associated with a subset of oral cancers, and yet, the natural history of oral HPV infection remains unexplored. The feasibility of studying oral HPV natural history was evaluated by collecting oral rinse samples on 2 occasions at a 6-month interval from 136 HIV-positive and 63 HIV-negative participants. Cervical vaginal lavage samples were concurrently collected for comparison. HPV genomic DNA was detected in oral and cervical samples by consensus primer PCR and type-specified for 37 HPV types. The six-month cumulative prevalence of oral HPV infection was significantly less than for cervical infection (p < 0.0001). HIV-positive women were more likely than HIV-negative women to have an oral (33 vs. 15%, p = 0.016) or cervical (78 vs. 51%, p < 0.001) infection detected. Oral HPV infections detected at baseline were as likely as cervical infections to persist to 6 months among HIV-negative (60% vs. 51%, p = 0.70) and HIV-positive (55% vs. 63%, p = 0.27) women. Factors that independently elevated odds for oral HPV persistence differed from cervical infection and included current smoking (OR = 8, 95% CI = 1.3-53), age above 44 years (OR = 20, 95% CI = 4.1-83), CD4 < 500 (OR = 6, 95% CI = 1.1-26), use of HAART therapy (OR = 12, 95% CI = 1.0-156), and time on HAART therapy (trend p = 0.04). The rate of oral HPV infections newly detected at follow-up was significantly lower than cervical infection among HIV-positive (p < 0.001) and HIV-negative women (p < 0.001). Our study not only demonstrates that it is feasible to study the natural history of oral HPV infection with oral rinse sampling, but also indicates that oral and cervical HPV natural history may differ.