The goal of the present PhD thesis is the modelling of formation and conversion of pollutants during the thermal utilization of process gases in an integrated steel plant by means of Cmputational Fluid Dynamics (CFD). The numerical calculations are done with the simulation software FLUENT from ANSYS Inc. to investigate the influence of operating conditions during a gas phase combustion. Steel plant heating systems represented the combustion reactors and the numerical calculations focused on the pollutant formation (especially NO and CO). Standard FLUENT sub models were used to predict the emission loads within a reasonable calculation time. This way of investigating the connection between pollution chemistry and operative process management during thermal utilization of metallurgical process gases was not done before. Especially the combustion of blast furnace gas was investigated. The results are used to realize specific operating conditions with regard to legal emission limits whereas the obtained information show extremely high relevance concerning a process efficiency optimization with respect to environmental responsibilities. A number of different mechanisms and phenomena such as the concentration of combustion intermediate species (radicals) play an important role in pollution chemistry. These components are formed and converted in the course of a combustion process. Moreover the pollution kinetics is strongly connected with the flow conditions in the reactor. Locally developed eddies, layers of species gas flow and recirculation areas respectively influence the reaction rate. These phenomena can lead to emission peaks due to an asymmetric gas flow distribution. Therefore it is absolutely necessary to consider the turbulence of flow as well as the heat release and heat transfer mechanisms e.g. radiative heat exchange with reactor walls and convection respectively. To evaluate the obtained CFD results small scale laboratory experiments were done investigating the combustion of a synthetically blast furnace gas mixture. Another evaluation step concerned a comparison of the calculated results with real plant emission data representing the highest quality evaluation method due to the fact that these reference data complies with the greatest possible extent to reality.