Nanoparticle gradient materials combine a concentration gradient of nanoparticles with a macroscopic matrix. This way, specific properties of nanoscale matter can be transferred to bulk materials. These materials have great potential for applications in optics, electronics, and sensors. However, it is challenging to monitor the formation of such gradient materials and prepare them in a controlled manner. In this study, we present a novel universal approach for the preparation of this material class using diffusion in an analytical ultracentrifuge. The nanoparticles diffuse into a molten thermoreversible polymer gel and the process is observed in real-time by measuring the particle concentrations along the length of the material to establish a systematic understanding of the gradient generation process. We extract the apparent diffusion coefficients using Fick's second law of diffusion and simulate the diffusion behavior of the particles. When the desired concentration gradient is achieved the polymer solution is cooled down to fix the concentration gradient in the formed gel phase and obtain a nanoparticle gradient material with the desired property gradient. Gradients of semiconductor nanoparticles with different sizes, fluorescent silica particles, and spherical superparamagnetic iron oxide nanoparticles are presented. This method can be used to produce tailored nanoparticle gradient materials with a broad range of physical properties in a simple and predictable way.