The stable silicon isotopic composition d30Si of silicic acid and of biogenic opal is used as a proxy for investigating the marine silicon cycle and silicic acid utilization by diatoms both at present and in the geological past. The marine biogeochemical and physical processes involved in determining the modern d30Si distribution have not been fully understood.Hence, the usage of d30Si as a proxy for reconstruction of the marine silicon cycle and paleoproductivity by diatoms is hampered. This work is aimed at providing a comprehensive view and systematic approaches for understanding the oceanic d30Si distribution and its controlling mechanisms under both present and the last glacial maximum (LGM) climate conditions. A coupled ocean (MPI-OM)-biogeochemical (HAMOCC5.1) model is applied to simulate the marine silicon cycle and the silicon isotopic fractionation processes during biogenic opal production and dissolution. In the present-day simulation, the surface d30Si increases along a Rayleigh type distillation curve during the utilization of silicic acid by diatoms, which demonstrates the primary control of biological fractionation on the surface d30Si distribution. The variations between the Rayleigh curves in different ocean basins, on the other hand, show the impact of physical transport of water on determination of the surface d30Si. In the deep ocean, our model captures a significant silicon isotopic gradient between the North Atlantic and the North Pacific. The advection related to the thermohaline circulation is thought to be the essential controlling factor of deep ocean d30Si. The model-data comparison implies that the usage of fractionation during biogenic opal dissolution as explanation to d30Si distribution is still speculative. The modeled silicic acid concentrations and d30Si show good agreement with the observations, when only fractionation during opal production is considered. The capability of the model to reproduce the large-scale modern oceanic d30Si distribution gives us confidence in simulating the d30Si during the LGM. In the LGM simulation, the extension of sea-ice cover in both the Northern and the Southern Hemisphere may cause a reduction of phytoplankton growth due to low light under the ice. The silicic acid utilization by diatoms especially around Antarctica is therefore inhibited, in line with reduced biogenic opal export fluxes. Our preliminary model results of the glacial Si isotopic composition agree with the interpretation from sediment core data that silicic acid utilization by diatoms in the Southern Ocean during the LGM was diminished relative to the present interglacial. The sensitivity of d30Si to glacial-interglacial ocean physical variations such as the strength of overturning circulation and ocean surface mixing is tested, using a simple seven-box model. The results indicate that the changes in global average d30Si due to the glacial-interglacial ocean physical variation may be of a similar magnitude as the total glacial-interglacial d30Si variation. The modeling approach is a valid and powerful tool of promoting a mechanistic understanding of the marine biogeochemical proxy d30Si. One important advantage over common interpretation of local field studies is that models calculate isotopic fractionation without application of Rayleigh or open system approximation. In addition, the sensitivity of d30Si to various biogeochemical and physical factors can be tested systematically.