This paper presents results obtained from the application of a computational fluid dynamics (CFD) code Fluent 6.3 to modeling of elevated pressure methane non-premixed sooting flames. The study focuses on comparing the two soot models available in the code for the prediction of the soot level in the flames. A standard k-ε model and Eddy Dissipation model are utilized for the representation of flow field and combustion of the flame being investigated. For performance comparison study, a single step soot model of Khan and Greeves and two-step soot model proposed by Tesner are tested. The results of calculations are compared with experimental data of methane sooting flame taken from literature. The results of the study show that a combination of the standard k-ε turbulence model and eddy dissipation model is capable of producing reasonable predictions of temperature both in axial and radial profiles; although further downstream of the flame over-predicted temperature is evidence. With regard to soot model performance study, it shows that the two-step model clearly performed far better than the single-step model in predicting the soot level in ethylene flame at both axial and radial profiles. With a modification in the constant α of the soot formation equation, the two-step model was capable of producing prediction of soot level closer to experimental data. In contrast, the single-step soot model produced very poor results, leading to a significant under-prediction of soot levels in both flames. Although the Tesner’s soot model is simpler than the current available models, this model is still capable of providing reasonable agreement with experimental data, allowing its application for the purpose of design and operation of an industrial combustion system.