Bisphenol A (BPA) is a harmful endocrine-disrupting compound. Accordingly, sensitive and rapid methods for BPA quantification are in high demand. In this study, a method was developed for relaxation signal amplification using a magnetic relaxation switch (MRS) biosensor based on the self-assembly of polystyrene microspheres and magnetic nanoparticles to detect BPA in food packaging materials and water samples. Aptamer-functionalized polystyrene microspheres (PS 1000 −Apt) and complementary DNAfunctionalized magnetic nanoparticles (MNP 20 −cDNA) self-assembled by forming a hybrid chain. The MNP 20 −cDNA can be captured by the larger size and increased aptamer loading of PS 1000 −Apt, and then it could be transformed from dispersion to aggregation. The aptamers could capture BPA with high affinity and selectivity, forming stable "PS 1000 −Apt−BPA" complexes and weakening MNP 20 −cDNA aggregation. As a result, an altered relaxation signal was obtained. Experimental studies confirmed that the proposed biosensor allowed accurate detection of BPA. Under optimized detection conditions, compared with a limit of detection (LOD) of 0.36 ng/mL for conventional MRS, a detection range (0.1−100 ng/mL) and a lower LOD (0.06 ng/mL) was obtained. Finally, the method was successfully applied in water and polycarbonate (PC) bottle samples for BPA for BPA detection. The PC bottles can be used to hold beverages, juices, and mineral water. However, they may contain a certain amount of BPA, which could be immersed in the solution, thereby affecting human health. The recoveries of spiked samples were 97.0−112.7%, and the relative standard deviation (RSD) values were 3.3−8.7%. These results were consistent with high-performance liquid chromatography studies. Collectively, the developed MRS sensor for relaxation signal amplification based on the self-assembly of polystyrene microspheres and magnetic nanoparticles (PS-MRS) could be a practical method for the ultrasensitive and accurate analysis of BPA in food packaging materials and water samples.