Quantum chemical descriptors (⑀ HOMO , ⑀ LUMO , absolute hardness, global softness, chemical potential, and electronegativity) and energy descriptors (Q min , ⌬H f 0 , E T , and E E ) based QSAR study of estrogen derivatives was made with the help of PM3 calculations on WinMOPAC 7.21 software. The observed RBA values of estrogens were taken from the literature. QSAR models were made using different quantum chemical and energy descriptors with the help of multiple linear regression analysis. Regression models indicate that absolute hardness in combination with different energy descriptors provide better correlation between observed relative binding affinity (RBA) and predicted relative binding affinity (PA). Regression models for other quantum chemical descriptors with energy descriptors are not as clear as in the case of absolute hardness. Hardness provides a better picture due to the maximum hardness principle and can be used as a QSAR model for predicting the biological activity of any compound.