Purpose The purpose of the present study is to determine if paternal or maternal history of diabetes mellitus (DM) and hypertension (HT) contributes to the prevalence and phenotype of polycystic ovary syndrome (PCOS). Methods We performed an epidemiologic study about PCOS from four districts in Beijing, China, between 2008 and 2009. Parental histories of DM and HT were collected, and the basic characteristics and serum indices of 123 PCOS patients and 718 non-PCOS controls were tested. Results The prevalence of a parental history of DM and HT was significantly higher in PCOS patients than non-PCOS women (17.1 % vs. 9.2 % and 42.3 % vs. 26.0 %, P<0.05, respectively). When paternal history was separated from maternal history, only a paternal history of DM and HT reached statistical significance between PCOS and non-PCOS patients (odds ratio (OR)=3.42, 95 % confidence interval (CI)=1.69-6.91; OR=2.50, 95 % CI=1.58-3.93, respectively). A paternal history of both DM and HT was significantly associated with sex hormone-binding globulin, fasting plasma glucose, and fasting insulin levels, the free androgen index, and the homeostatic model assessment-insulin resistance in PCOS patients (P<0.05 for all). There was no independent association between maternal history and the clinical or biochemical phenotype of PCOS. Conclusions PCOS patients with a positive paternal history of both DM and HT have an adverse endocrine and metabolic profile. A paternal history of DM and HT poses a risk to PCOS.