We present an application of fluorescence recovery after photobleaching (FRAP) to measure the size of the individual nanoparticles in binary systems. The presence of nanoparticles with varying sizes was successfully demonstrated using a straightforward biexponential model and their sizes were accurately determined. Furthermore, we have demonstrated the benefits of preprocessing the data using a simple machine learning algorithm based on the gradient boosting machine and fitting the resulting curves to a triexponential model. This approach allows the accurate recovery of the sizes of each of the three components in a binary particle system, namely, the 6 nm LUDOX HS40, 11 nm LUDOX AS40, and the free R6G labeling dye. Lastly, it has been demonstrated using molecular dynamics simulations that R6G adsorption to silica nanoparticles (SNPs) is indeed size-dependent, with larger constructs as the preferred target because of their higher charge and smaller curvature. The theoretical and experimental results were therefore consistent with one another.