The advantages of closed, loop over open loop control systems are generally recognised. However, existing engine management systems implement most control functions in open loop because suitable feedback sensors are not available. Even for so-called closed loop air fuel ratio controllers, shortcomings of the exhaust gas oxygen (EGO) sensor limit the potential effectiveness of closed loop control. A more direct measure of the combustion process, such as cylinder pressure, can yield sufficient information for the closed loop operation of many of the combustion control functions; this paper presents the results of a prediction algorithm which can derive a variety of feedback signals from cylinder pressure.Cylinder pressure, together with several combustion variables, including air-fuel ratio, exhaust gas recirculation rate, and NOx HC, CO and CO2 emissions were measured at various operating points. The combustion variables were assumed to be a function of cylinder pressure and the parameters of this function were identified from the experimental data. Predictions of combustion variables were generated from individual cycles of cylinder pressure data using this function. These predictions show good correlation with the measured values and could form the basis of a full authority, high performance closed loop combustion controller. Potential benefits include reduced exhaust gas emissions, improved fuel economy and reduced roughness.The technique can also be applied to data from other combustion sensors and combustion parameters. Thus other combustion sensors and additional combustion control functions could be considered for a practical engine management system.