In this thesis, the large-scale thermohaline circulation in the Late Permian was studied. A steady state circulation was obtained by running MITgcm with a 2 times 2 degrees spherical-polar grid, and solving the nonlinear equations of motion until an equilibrium state was obtained. The results were compared to some linear theories on the ocean circulation, and a similar model study for a square basin ocean. The goal was to discuss the relevancy of these simplifications to the Late Permian ocean. Additionally, the results were discussed in terms of some previous numerical and proxy-based studies. The objective was to illustrate how the ocean circulation in this period could have contributed to the anoxic and euxinic conditions observed on the Permo-Triassic boundary, in relation the End Permian extinction, and to explore the uncertainties related to plaeoclimatic studies, both numerical and paleogeological. Such comparison studies can be useful in order to explore the causes for oxygen depletion in the Late Permian ocean, and separate out the causes for local extinction rates. The conditions on Permo-Triassic boundary are also highly relevant in order to understand the current climate changes, as rapid increase of atmospheric CO2 is thought to be the main trigger of the End Permian extinction. It was shown that both the linear models and the square basin case holds some similarities to the Late Permian ocean, but neither is a satisfactory simplification. Regarding the previous numerical and proxy-based studies, it was found that the circulation is highly sensitive to bathymetry, particularly in highly studied areas. Also, localities where anoxic conditions is seen in the proxy data, do not necessarily coincide with areas of stagnation. The results presented in this thesis consists only of a snap shot of the ocean circulation prior to the extinction event. Thus it can not be directly compared to the situation during and after the extinction. The aim is to illustrate how pairing numerical models and geological data can be advantageous when studying this particular event.