Autoregulation is a mechanism necessary to maintain an approximately constant blood flow rate in the microcirculation when acute changes in systemic pressure occur. Failure of autoregulation in the retina has been associated with various diseases, including glaucoma. In this work, we propose an initial attempt to model autoregulation in a 3D network of retinal arteries. The blood flow is modeled with the time-dependent Stokes equations. The arterial wall model includes the endothelium and the smooth muscle fibers. Various simplifying assumptions lead to a fluid-structure model where the structural part appears as a boundary condition for the fluid. The numerical simulations are performed on a patient-specific network of 25 segments of retinal arteries located in the inferior temporal quadrant. The results are first compared with experimental data for a given value of perfusion pressure. Then, to assess the autoregula-tion mechanism, flow rate-pressure curves are simulated with various perfusion pressures. The results obtained with the proposed 3D fluid-structure model are in good agreement with experimental data and 0D models.