The heavy-duty diesel (HDD) and light-duty diesel (LDD) have been required to meet diesel emission standards that will become mandatory worldwide in the next few years (i.e., Euro 6 for Europe and Tier 2 for USA light-duty vehicles). Currently, selective catalytic reduction (SCR) of NOx, with urea (commercially known as AdBlue) as a reducing agent, has been employed in the NOx emission control of HDD. The commercial catalyst used is based on V2O5-WO3-TiO2, but the V 2O5 must be replaced because of its toxicity. Catalysts based on metal-exchanged zeolite have been widely studied due to their high catalytic activity and stability for SCR-NOx. The catalysts, based on the ZSM-5 zeolite (CuZSM5 and FeZSM5) and Cuban natural mordenite (CuMORD), have been prepared by a conventional ion-exchange method and tested in the SCR of NOx with ammonia in presence of water vapor and SO2. This paper deals with these experimental results and will exploit a neural network based on the approach to predict the NO conversion efficiency of a commercial catalyst and implement the derived mathematical function into a numerical model for diesel road vehicle simulation. Light-duty and heavy-duty vehicles are going to be simulated in both congested and noncongested roads to analyze the SCR system behavior in such conditions.