A computational fluid dynamic (CFD) model of a transpiring wall reactor for supercritical water oxidation has been implemented and solved using the commercial software Fluent 6.3. Model results have been validated by comparison with experimental temperature profiles, and the influence of model simplifications on the accuracy of the results has been discussed. It has been found that the model presents some inaccuracies in the calculation of the maximum reaction temperature, which have been attributed to the simplifications in the calculation of thermal properties of the fluids imposed by limitations of the software. The effect of different process parameters has been studied, including transpiring water flow ratio, transpiring water temperature, flow rate and composition of the oxidant (pure oxygen or air) on different indicators of reactor performance such as temperature and composition contours, flow path lines and effluent compositions. The behaviour of three different transpiring wall designs comprising different sections of porous and non-porous materials has also been investigated. It has been found that making the upper cup of the reactor of a porous material has a very little influence on reactor performance, and therefore it is preferable to build this part of the reaction chamber using a non-porous, more durable material. On the other hand, the maximum reaction temperature in the reaction chamber can be increased if the upper part of the reaction chamber is made with a non-porous material, which is favourable for treating diluted feeds. Model results complement previous information obtained by experiments and by simplified models. (C) 2010 Elsevier B.V. All rights reserved.