In the last two decades, piston engine specifications have deeply evolved. Indeed, new challenges nowadays concern the reduction of pollutant emissions (EURO regulations) and CO2 emissions. To satisfy these new requirements, powertrains have become very complex systems including a large number of high technology components (high pressure injectors, turbocharger, Exhaust Gas Recirculation (EGR) loop, after-treatment devices...). In this context, the engine control plays a major role in the development and the optimization of powertrains.Few years ago, engine control strategies were mainly defined by experiments on engine test benches. This approach is not adapted to the complexity of future engines: on the one hand, tests are too expensive and on the other hand, they do not give much information to understand interactions between components. Today, a promising alternative to tests may be the use of 0D/1D simulation tools. These methods have been widely used in the past ten years and allow building engine control algorithms. However, they are generally based on empirical models and often suffer from a lack of predictivity. A solution for extending the range of application of the system simulation consists in developing more physical models based on the 3D calculations experience. This way has been recently followed at IFP Energies nouvelles, leading to the development and implementation of several libraries dedicated to powertrains (IFP-Engine, IFP-Exhaust) and drivetrains (IFP-Drive) in the AMESim simulation software.The work we propose here is coherent with the approach chosen at IFP Energies nouvelles and aims at developing a phenomenological model for the pollutants emissions in a combustion chamber of a piston engine.For this purpose, a model able to take into account multiple injections, conventional diesel mode and Homogeneous Charge Compression Ignition (HCCI) mode have been developed modeling physical phenomena like: - spray atomization and the liquid phase penetration, - vaporization deduced from a characteristic time, - presumed mixing distribution computed with a β-function where the mixture fraction variance equation is obtained from the integration of turbulence, inlet/outlet spray zone mass and evaporation, and - the auto-ignition and diffusive combustion regimes determined by Flame Propagation of ILDM (FPI/Intrinsic Low-Dimensional Manifolds) tabulation combined with turbulence regime by a Presumed Conditional Moments (PCM) method.In this paper, we propose a modeling of NOx based on tabulations of a characteristic time and NO mass fraction equilibria. After a brief introduction of the combustion model, the NOx model is theoretically described and results on a diesel conventional engine are presented, compared to experimental data.