The use of numerical simulations in the development processes of engineering products has been more frequent, since it enables prediction of premature failures and study of new promising concepts. In industry, numerical simulation has the function of reducing the necessary number of validation tests prior to spending resources on alternatives with lower likelihood of success. The internal combustion Diesel engine plays an important role in Brazil, since they are used extensively in automotive applications and commercial cargo transportation, mainly due to their relevant advantage in fuel consumption and reliability. In this case, the most critical pollutants are oxides of nitrogen (NOx) and particulate matter (PM) or soot. The reduction of their levels without affecting the engine performance is not a simple task. This paper presents a methodology for guiding the combustion analysis by the prediction of NOx emissions and soot using numerical simulation. The methodology includes the use of 1D flow analysis using GT-Power and a three dimensional (3-D) computational fluid dynamics (CFD) to model flow inside the cylinder (using KIVA code), including models for turbulence, jet break-up process and models to predict emissions using kinetic chemistry.Injector design parameters were varied in order to evaluate the fidelity of the model through the qualitative trends found in tests. The effects of variations in engine load and fraction of dilution were also considered. The models were used to evaluate the sensitivity to variation of injector parameters.The results show good success in determining the qualitative trends in the simulated emissions values, especially regarding NOx estimation, for which the model correctly predicted the emissions trends in 93% of the cases.