In the chronic ocular disease wet age-related macular degeneration (AMD), current therapies rely on intravitreal (back of the eye) injections of anti-vascular endothelial growth factor agents every four to six weeks to maintain therapeutic drug levels. Controlled-release drug delivery systems offer a promising alternative by extending drug release to reduce injection frequency. This study developed a continuum diffusion model for drug transport through porous polymeric microcapsules, implemented using the finite element method in COMSOL Multiphysics. The case study focused on cylindrical annular microcapsules fabricated with either a single polycaprolactone (PCL) layer or a bi-layered chitosan-PCL structure, tested at three fabrication-induced porosity levels. Bovine serum albumin and bevacizumab were used as model drugs. Parameter estimation was performed using published release data [1]. The model reproduced experimental release profiles across formulations and identified key transport parameters governing release dynamics, including porosity, tortuosity, and mass transfer rates. Design exploration revealed that polymer thickness was the dominant variable controlling release, while addition of the chitosan layer moderated the initial burst and extended therapeutic delivery. The design exploration simulations from the high fidelity finite element simulations were processed to approximate the release rates by regression to two-term exponential models. These models predicted the effective time or duration that the release rate stayed above the therapeutic threshold. Then Gaussian process regression (GPR) was used to map the correspondence between effective time and the design variables of polymer layer thickness and PCL:chitosan ratio across the design space. Bayesian optimization on the GPR surrogate model was used to determine the optimal combination of design parameters to maximize the effective time. This framework demonstrates how computational modeling can reduce experimental burden, guide design optimization, and support the development of long-acting intravitreal drug delivery systems to treat wet AMD by linking drug release kinetics to design variables. References: [1] Jiang P., Chaparro F. J., Cuddington C. T., Palmer A. F., Ohr M. P., Lannutti J. J., Swindle-Reilly K. E., Injectable biodegradable bi-layered capsule for sustained delivery of bevaciin treating wet age-related macular degeneration, J Controlled Release, Vol. 320, pp. 442, 2020.