One of the most vital factors in combustion control is Air-to-Fuel Ratio (AFR) control and estimation. Various parameters such as intake air mass flow rate, fuel injection quantity and (Exhaust Gas Recirculation) EGR flow rate are affecting the AFR magnitude. In this study a detailed mathematical, nonlinear and control oriented model of dynamic processes of turbocharged diesel engines is presented for the purpose of AFR control. It is an attempt to describe the two main subsystems of compression ignition engines called intake manifold and fuel injection systems.The intake model itself consists of compressor, intercooler, intake manifold, EGR circuit, valve section and cylinder subsystems. This model has been developed for study of the air intake system in a turbocharged diesel engine, using physical equations. Experimental data has also been used to adjust the model. The output variable of this submodel (air mass flow) is a significant parameter in calculations of engine power, emission and AFR control. Of the methods of assisted aspiration, the exhaust gas driven turbocharger is by far the most widely used. Varying the rate of gas flow through the turbine by means of Variable Geometry Turbine (VGT) is a method by which the exhaust gas flow through turbine can be limited at high engine speeds. Therefore, turbocompound in this model is considered to have VGT.Common Rail Injection (CRI) that is a fully flexible fuel injection system in which quantity, timing and pressure of injection are controllable (based on electronic control) separately is chosen in this study. Fuel injection system is comprised of solenoid, working chamber and needle subsystems. A one-dimensional, transient and compressible flow model of CRI, based on Kirchhoff's law, mass and momentum conservation equations and the equilibrium of forces is derived.AFR control of diesel engine is performed make use of fuzzy logic methodology. Fuzzy logic is a control structure which emulates the way humans arrive at decisions, given a certain set of circumstances. This method improves the control and simplifies the development process. The fuzzy logic controller achieves this by considerably reduced complexity and lack of tedious mathematical derivation associated with modern control theory methods.All above-mentioned models are programmed in Matlab/Simulink software environment. The simulation results are then compared with available data to check the accuracy of the model.