Concern for combustion generated pollution of the environment has led to intense research on the subject. Some of it has already spawned significant technical innovations, particularly in the industrial sector. Still more stringent control of pollutant emissions is being demanded by the public, however, and therefore continued fundamental research on pollution control aspects of the combustion process is being supported in the industrial countries. The aim of this thesis, soot modeling in non-pre mixed flame. To achieve this goal, first model has been identified related to the soot and species of influence in formation and soot oxidation have been studied. Soot is the most popular pollutant. Different stage of Soot formation includes five stages: 1- Inception 2- level expansion 3- Accumulation of particles 4- Coagulation 5- oxidation. Stage of spermatogenesis for the most important and most complex stage of soot formation is different reaction has been suggested that among the most important formation related to the first reaction is a hydrocarbon ring.C 2 H 2 Species are most effective in formation of soot and various studies on the effect of these stresses is. O 2 and OH In most sources as the main oxidants are used. O in some sources as oxidizing soot is considered. The reaction mechanisms of reactive species are generally fall into three categories. The first one deals with extensive mechanisms, where many species and reactions are considered for the chemical kinetics. The second category makes that use of the reduced mechanisms to mimic the prevailing chemical kinetics of the reactions. The third category of reaction mechanisms are global reactions, mostly one- and two-step reactions. Two well-known comprehensive mechanisms are compared for the prediction of an opposed-jet diffusion flame. Results show that the GRI-MECH3 mechanism is superior over the Konnov mechanism in terms of computational time. Models presented two groups a single equation and two equations can be divided into. The main difference in the model presented can be in coefficients of the model. Each of these models has been applied in the CFD code.Turbulence model building in flames from the standard k-? method is used. The CFD code using the detection values for numerical modeling of this flame is used. The results show that the single equation model results for how turbulence and how laminar flame for the flame to confuse not offer. A result of two-equation model for flame confused does not have a good accuracy. Key Words: Soot, Diffusion flame, Chemical mechanism, Combustion, Flamelet, CFD