Objective To determine endometrial cancer (EC) risk according to family cancer history, including assessment by degree of relatedness, type of and age at cancer diagnosis of relatives. Methods Self-reported family cancer history was available for 1353 EC patients and 628 controls. Logistic regression was used to quantify the association between EC and cancer diagnosis in ≥ 1 first or second degree relative, and to assess whether level of risk differed by degree of relationship and/or relative's age at diagnosis. Risk was also evaluated for family history of up to three cancers from known familial syndromes (Lynch, Cowden, hereditary breast and ovarian cancer) overall, by histological subtype and, for a subset of 678 patients, by EC tumor mismatch repair (MMR) gene expression. Results Report of EC in ≥ 1 first- or second-degree relative was associated with significantly increased risk of EC (P = 3.8 × 10− 7), independent of lifestyle risk factors. There was a trend in increasing EC risk with closer relatedness and younger age at EC diagnosis in relatives (PTrend = 4.43 × 10− 6), and with increasing numbers of Lynch cancers in relatives (PTrend ≤ 0.0001). EC risk associated with family history did not differ by proband tumor MMR status, or histological subtype. Reported EC in first- or second-degree relatives remained associated with EC risk after conservative correction for potential misreported family history (OR 2.0; 95% CI, 1.24–3.37, P = 0.004). Conclusion The strongest predictor of EC risk was closer relatedness and younger EC diagnosis age in ≥ 1 relative. Associations remained significant irrespective of proband MMR status, and after excluding MMR pathogenic variant carriers, indicating that Lynch syndrome genes do not fully explain familial EC risk.