Overloading and climate change are often problems in pavement structures. For this reason, hard asphalt binders have high softening points, are elastic, and have good adhesion, which is needed to improve pavement performance. Asphalt binder performance can be enhanced by adding additives such as natural rubber or natural-rubber-modified asphalt. However, natural-rubber-modified asphalt shows poor storage stability problems. This is due to differences in density and viscosity between the constituent components of natural-rubber-modified asphalt. This study examines the phase separation mechanism in technically specified natural rubber (TSNR) modified asphalt. Prediction of the optimum storage length of modified asphalt before phase separation occurs, using a combined incompressible Navier–Stokes and phase field model and carried out with COMSOL Multiphysics software version 5.5. Experimental validation was conducted at TSNR levels of 8, 10, and 12% at 160 °C for 48 h, with and without sulfur. The simulation showed that the asphalt modified with TSNR experienced phase separation after 12 h of storage at 160 °C under conditions without stirring. This aligns with the experimental results, which showed phase separation at 160 °C after 48 h. Adding sulfur additives did not have much effect on improving storage stability. The combined incompressible Navier–Stokes and phase field model accurately describes the phase separation in TSNR-modified asphalt. The results of this research recommend that the industry store natural-rubber-modified asphalt in a constantly stirred condition to prevent phase separation of modified asphalt. In addition, the results of this research help the industry predict or increase the homogeneity of polymer-modified asphalt production and save time and costs.