Background  The presence of a wide range of tests of ovarian reserve suggests that no single test provides a sufficiently accurate result. Many tests are used without reference to an evidence base. So far, individual studies conducted on these tests are too small to give precise estimates of prognostic accuracy. Objectives  To systematically assess the accuracy of the available tests of ovarian reserve in terms of prediction of fertility outcomes. Search strategy  The search will be conducted using the name of the respective index test being studied (as listed on the MESH database), if more than 2000 citations are listed, ‘ovary’ and or ‘ovarian’, ‘fertility’ and or ‘reserve’ will be combined with the original search term as required. Studies of the accuracy of tests of ovarian reserve will be obtained without language restrictions from 1980 to 2005 using the following electronic databases and Ovid software: MEDLINE, EMBASE, PUBmed, Biological extracts, Pascal, Cochrane Library (CDSR, DARE, CCTR, HTA), Best Evidence databases, SCISEARCH, Conference Proceedings (ISI Proceedings, Healthstar, Current Contents, Science Citation Index, Cancerlit and Econlit and NHS Economic Evaluation database. The National Research Register, the Medical Research Council’s Clinical Trials Register, MEDION, DARE, and the US Clinical Trials register. Selection criteria  Studies will be selected if accuracy of tests are compared with a reference standard and include data that can be abstracted into a two‐by‐two table to calculate sensitivity and specificity. The studies to be included in this review will examine one of the following index ‘tests’ within a study population of women undergoing assisted reproductive technology: Clinical variables—age, history of cancelled cycles. Basal blood tests—follicle‐stimulating hormone (FSH), lutenising hormone (LH), FSH:LH ratios, estradiol (E2), inhibin A and B, progesterone (P4), P4:E2 ratios, antimullerian hormone, testosterone, vascular endothelial growth factor, insulin‐like growth factor‐1:insulin‐like growth factor binding protein‐1 ratios. Dynamic tests—clomiphene citrate challenge test, gonadotropin analogue stimulating test, exogenous FSH ovarian reserve test. Ultrasound tests—antral follicle count, ovarian volume, ovarian stromal peak systolic velocity, including waveform and pulsatility index, ovarian follicular vascularity. Histology—ovarian biopsy. Data collection and analysis  Two independent reviewers will perform quality assessment and data extraction. Prognostic accuracy will be determined by calculating positive and negative likelihood ratios for the following outcomes or reference standards: live birth, ongoing pregnancy, clinical pregnancy, biochemical pregnancy, embryos available for transfer, eggs obtained at oocyte retrieval, cycles cancelled prior to oocyte retrieval. Main results and conclusions  N/A.