The rise of nanofiltration technologies holds great promise for creating more effective and affordable techniques aiming to remove undesirable pollutants from wastewaters. Despite nanofiltration’s promising potential in removing antineoplastic drugs from liquid matrices, the limited information on this topic makes it important to estimate the rejection rates for a larger number of compounds, particularly the emerging ones, in order to preview the nanofiltration performance. Aiming to have preliminary estimations of the rejection rates of antineoplastic drugs by nanofiltration, 54 antineoplastic drugs were studied in 5 nanofiltration membranes (Desal 5DK, Desal HL, Trisep TS-80, NF270, and NF50), using a quantitative structure-activity relationship (QSAR) model. While this methodology provides useful and reliable predictions of the rejections of compounds by nanofiltration, particularly for hydrophilic and neutral compounds, it is important to note that QSAR results should always be corroborated by experimental assays, as predictions were confirmed to have their limitations (especially for hydrophobic and charged compounds). Out of the 54 studied antineoplastic drugs, 29 were predicted to have a rejection that could go up to 100%, independent of the membrane used. Nonetheless, there were 2 antineoplastic drugs, fluorouracil and thiotepa, for which negligible removals were obtained (<21%). This study’s findings may contribute (i) to the selection of the most appropriate nanofiltration membranes for removing antineoplastic drugs from wastewaters and (ii) to assist in the design of effective treatment approaches for their removal.