Abstract Background No screening programs are available for ovarian or endometrial cancer. One reason for this is the low incidence of the conditions, resulting in low positive predictive values for tests, which are not very specific. One way of addressing this problem might be to use risk factors to define subpopulations with a higher incidence. The aim of this study was to investigate the extent to which a medical history of endometriosis can serve as a risk factor for ovarian or endometrial cancer. Methods In a hospital-based case–control analysis, the cases represented patients with endometrial or ovarian cancer who were participating in studies aimed at assessing the risk for these diseases. The controls were women between the age of 40 and 85 who were invited to take part via a newspaper advertisement. A total of 289 cases and 1016 controls were included. Using logistic regression models, it was tested whether self-reported endometriosis is a predictor of case–control status in addition to age, body mass index (BMI), number of pregnancies and previous oral contraceptive (OC) use. Results Endometriosis was reported in 2.1 % of the controls ( n = 21) and 4.8 % of the cases ( n = 14). Endometriosis was a relevant predictor for case–control status in addition to other predictive factors (OR 2.63; 95 % CI, 1.28 to 5.41). Conclusion This case–control study found that self-reported endometriosis may be a risk factor for endometrial or ovarian cancer in women between 40 and 85 years. There have been very few studies addressing this issue, and incorporating it into a clinical prediction model would require a more precise characterization of the risk factor of endometriosis.