Purpose To evaluate basal testosterone (T) levels in women undergoing in vitro fertilization (IVF) cycles and examine the association between basal T levels and ovarian response or IVF pregnancy outcome. Methods We retrospectively analyzed 1413 infertile Chinese women undergoing their first IVF treatment at our institution's reproductive center from March 2011 to May 2013. The basal testosterone (T) levels in women undergoing in vitro fertilization (IVF) and the relationship between basal T levels and ovarian response or IVF pregnancy outcome were determined. These patients did not have polycystic ovary syndrome (PCOS) or endometriosis, and were treated with a long luteal down-regulation protocol. Subjects were divided into 2 groups according to basal testosterone (T) levels: Group 1, basal T values <20 ng/dl (n=473), and Group 2, basal T values >20 ng/dl (n=940). We evaluated the association of basal T levels with ovarian response and IVF outcome in the two groups. Results In this study, BMI, basal follicle-stimulating hormone (FSH) levels, basal luteinizing hormone (LH) levels, antral follicle count (AFC), days of stimulation, total gonadotrophin dose, basal FSH/LH ratio, and the number of follicles >14 mm were significantly different (P<0.05) between the two groups. Basal T level positively correlated with ovarian reserve function, number of follicles >14 mm on human chorionic gonadotrophin (HCG) day, and total gonadotropin dose. However, basal T levels play no role in predicting IVF pregnancy outcome. Conclusion Basal T level can be used as a good predictor for ovarian response and the number of large follicles on HCG day. Additionally, we may use basal T level as a marker to predict FSH dosage. In general women, lower level of T might relate with potential poor ovarian response. However, based on our data, basal T levels do not predict pregnancy outcome.