In modern Diesel engine control strategies the guideline is to perform an efficient combustion control, mainly due to the increasing request to reduce pollutant emissions. Innovative control algorithms for optimal combustion positioning require the on-board evaluation of a large number of quantities. In order to perform closed-loop combustion control, one of the most important parameters to estimate on-board is MFB50, i.e. the angular position in which 50% of fuel mass burned within an engine cycle is reached. Furthermore, MFB50 allows determining the kind of combustion that takes place in the combustion chamber, therefore knowing such quantity is crucial for newly developed low temperature combustion applications (such as HCCI, HCLI, distinguished by very low NOx emissions). The aim of this work is to develop a virtual combustion sensor, that provides MFB50 estimated value as a function of quantities that can be monitored real-time by the Electronic Control Unit (ECU).Modern technologies for Common Rail Multi-Jet Diesel engines allow designing injection patterns with many degrees of freedom, due to the large number of tunable injection parameters (such as rail pressure, start and duration of each injection….). First, this paper describes a model of the combustion process developed in order to evaluate the energy release within the cylinder. A zero-dimensional approach based on the Wiebe function has been chosen, because it allows obtaining a model which is accurate enough for the analysis at a low computational cost. Once the combustion model has been developed it can be used to determine MFB50. The second section of this paper describes the existing correlations between the injection parameters and the identified Wiebe parameters. These correlations can be used for heat release and MFB50 on board estimation.Experimental tests have been performed running a turbocharged Common Rail Multi-Jet Diesel engine (with up to 4 injections within the same engine cycle) in order to determine the accuracy of the methodology.The described approach allows evaluating MFB50 as a function of the injection parameters as well as other quantities that can be monitored real-time by the ECU. Additional sensors are not necessary for this methodology, therefore it requires no extra cost.