Objective: The aim of this study was to investigate whether anti-M$ddot{u}$llerian hormone (AMH) levels could be predict ovarian poor/hyper response and IVF cycle outcome. Methods: Between May 2010 and January 2011, serum AMH levels were evaluated with retrospective analysis. Three hundred seventy infertile women undergoing 461 IVF cycles between the ages of 20 and 42 were studied. We defined the poor response as the number of oocytes retrieved was equal or less than 3, and the hyper response as more than 25 oocytes retrieved. Serum AMH was measured by commercial enzymelinked immunoassay. Results: The number of oocytes retrieved was more correlated with the serum AMH level (r=0.781, $p$ <0.01) than serum FSH (r=-0.412, $p$ <0.01). The cut-off value of serum AMH levels for poor response was 1.05 ng/mL (receiver operating characteristic [ROC] curves/area under the curve [AUC], $ROC_{AUC}$=0.85, sensitivity 74%, specificity 87%). Hyper response cut-off value was 3.55 ng/mL $ROC_{AUC}$=0.91, sensitivity 94%, specificity 81%). When the study group was divided according to the serum AMH levels (low: <1.05 ng/mL, middle: 1.05 ng/mL - 3.55 ng/mL, high: >3.55 ng/mL), the groups showed no statistical differences in mature oocyte rates (71.6% vs. 76.5% vs. 74.8%) or fertilization rates (76.9% vs. 76.6% vs. 73.8%), but showed significant differences in clinical pregnancy rates (21.7% vs. 24.1% vs. 40.8%, $p$=0.017). Conclusion: The serum AMH level can be used to predict the number of oocytes retrieved in patients, distinguishing poor and high responders.