Rubbery polymer networks, including elastomers and hydrogels, are increasingly employed in advanced applications such as biomedical implants, drug delivery systems, and smart sensors and actuators. Their macroscopic mechanical properties, such as stiffness, strength, and stretchability, are largely governed by network-level features, including polymer chain length distribution, crosslink density, and network heterogeneities and topological defects. Consequently, recent material design strategies have focused on tailoring network architecture to achieve enhanced mechanical performance. Despite these efforts, predictive tools that directly connect molecular- and network-scale parameters to macroscopic mechanical response remain scarce. Atomistic simulations are computationally prohibitive at relevant length and time scales, while continuum models often rely on phenomenological fitting parameters. In this talk, I will present our recent progress toward a predictive multiscale modeling framework for soft rubbery networks that incorporates microscale physics at the single-chain and network levels.