PURPOSE Cervical cancer screening is one of the strategies to prevent the disease among women at risk. Human papillomavirus (HPV) DNA testing is increasingly used as the cervical cancer screening method because of its high sensitivity. Self-collection of cervical specimens has the potential to improve participation. However, there is only limited information on comparison between self-collected and provider-collected samples with regard to detection of high-risk HPV using the careHPV method. The study aimed to compare HPV detection by careHPV in self-collected and provider-collected cervical samples and to assess the acceptability of self-collection techniques. MATERIAL AND METHODS Women attending cervical cancer screening clinics at Ocean Road Cancer Institute, Kilimanjaro Christian Medical Centre or Mawenzi Hospital in Tanzania were included in the study. They underwent a face-to-face interview, HIV testing, and collected a self-sample using Evalyn Brush. Subsequently, they had a cervical sample taken by a health provider. Both samples were tested for high-risk HPV DNA using careHPV. RESULTS Overall, 464 women participated in the study. The high-risk HPV prevalence was 19.0% (95% CI, 15.6 to 22.9) in the health provider samples, but lower (13.8%; 95% CI, 10.9 to 17.3) in the self-collected samples. There was a good overall agreement 90.5% (95% CI, 87.5 to 93.0) and concordance (κ = 0.66; 95% CI, 0.56 to 0.75) between the two sets of samples. Sensitivity and specificity were 61.4% (95% CI, 50.4 to 71.6) and 97.3% (95% CI, 95.2 to 98.7), respectively, varying with age. Most women preferred self-collection (79.8%). CONCLUSION Overall, self-sampling seems to be a reliable alternative to health-provider collection and is acceptable to the majority of women. However, instructions on proper procedures for sample collection to the women are important.