The homologation process of the internal combustion engines (ICE) is becoming more difficult with the passing of time. Coupled with the need for massive investments in new technologies, such as hybrid or hydrogen powertrains or 100% electrical vehicles, the economy in the traditional automotive manufacturing is stretched to its limits. There is stringent need to limit as much as possible the development cost of the new ICEs set for the Euro7 emissions standards. The last 15 years have seen a giant leap forward in computer-aided design (CAD) and computer-aided engineering (CAE) in the imagining and validation of different structures. This software can now reproduce with outstanding precision the aeraulic phenomena in the air intake, in the combustion chamber and in the exhaust of ICEs. The aim of this paper is to present a general overview on system modeling in ICEs, with an emphasis on the use of predictive and non-predictive combustions models. Thus, a special attention was given to the SITurb predictive combustion method, which was employed to build a virtual engine. By modifying the models' parameters, the spark timing was corelated with the values issued from the engine test bench. Some of the tuning activities usually done on the engine test bench could be done on the SITurb model, with similar results to the tests done on physical test supports. Using this approach could lead to economic advantages, in terms of time and material resources employed in the development of the ICEs.