Three-dimensional quantitative structure-activity relationship (3D-QSAR) methods were innovatively introduced into the structure-performance study of flotation collectors using topomer comparative molecular field analysis (Topomer CoMFA) and comparative molecular similarity indices analysis (CoMSIA). A total of eighteen cupferron derivatives was used to build up the 3D-QSAR models. The resultant Topomer CoMFA and CoMSIA models are both statistically significant. The Topomer CoMFA model yielded the non-cross-validated correlation coefficient (r(2)) and the leave-one-out correlation coefficient (q(2)) of 0.931 and 0.707, respectively. The corresponding r(2) and q(2) of the CoMSIA model were 0.816 and 0.651, respectively. The theoretically predicted separation efficiencies are in excellent agreement with the experimentally measured values, validating the high predictive ability of the 3D-QSAR models. Valuable information concerning the hydrophobic, steric and electrostatic properties of substituted cupferrons has been obtained and analysed in detail. The 3D-QSAR contour maps indicate that the aliphatic chains with significant hydrophobicity together with electron rich groups at 4-position of the benzene are of utmost importance for the flotation performance of substituted cupferrons. The addition of the electron-withdrawing group to the 2, 3-positions of the benzene ring is undesirable for enhancing the separation efficiencies. Structural modification of substituted cupferrons has been evaluated and novel cupferron derivatives with theoretically improved separation efficiency have been proposed. Despite the fact that 3D-QSAR methods have matured into a wide variety of applications such as drug discovery, catalyst science, environmental science, pesticide science, computational toxicology and material science, this work is the first attempt of introducing 3D-QSAR into the structure-activity relationship of flotation collectors. The generated 3D-QSAR models in this work could provide a new perspective on the structure-activity relationship of flotation collectors and potentially hold great promise for the rational design, selection and prediction of untested flotation collectors through virtual screening of chemical database.