Breast cancer is a cancer caused by uncontrolled cell growth at breast tissue. One of the most common triggers of breast cancer is overexpression of estrogen receptor alpha (ERα). This research's goal is to test the ability of coumestrol as the ligand of ERα with in silico method and to discover coumestrol's binding pose inside the ERα's binding pocket. Coumestrol's ability as ERα's ligand was tested using structure-based virtual screening (SVBS) method by Setiawati et al. (2014) that had been modified by Istyastono (2015). Results analysis was done using decision tree generated from recursive partition and regression tree method (RPART). If coumestrol is a ligand based on decision tree, it is concluded that coumestrol is active as ligand of ERα. At the end of analysis, coumestrol's pose inside ERα's binding pocket was visualized using MacPyMol. From the test acknowledged that the smallest ChemPLP value of coumestrol's pose was -83.1487. Coumestrol interacts with GLY420, ARG394, and GLU353 using hydrogen bonds. However, coumestrol were perceived as decoy according to decision tree. Hence, coumestrol could not be recognized as ERα's ligand by the protocol. Therefore, development of proper protocol to indentify ligand for ERα is required.