Automotive engines and control systems are more and more sophisticated due to increasingly restrictive environmental regulations. Particularly in both diesel and SI lean-burn engines NOx emissions are the key pollutants to deal with and sophisticated Engine Management System (EMS) strategies and after-treatment devices have to be applied. In this context, the in-cylinder oxygen mass fraction plays a key-role due its direct influence on the NOx formation mechanism. Real-time estimation of the intake O₂ charge enhances the NOx prediction during engine transients, suitable for both dynamic adjustments of EMS strategies and management of aftertreatment devices.The paper focuses on the development and experimental validation of a real-time estimator of O₂ concentration in the intake manifold of an automotive common-rail diesel engine, equipped with turbocharger and EGR system. The paper analyzes the air intake process and the influence of the exhaust gas recirculation system based on a mean value modeling approach. All variables required are available at EMS level, thus allowing an on-board implementation without extra costs for additional sensors.The accuracy of the developed estimator is assessed by comparing simulated and experimental trajectories of O₂ concentration, measured by a Universal Exhaust Gas Oxygen (UEGO) sensor located in the intake manifold. The experimental tests were carried out at the test bench, imposing severe engine transients. The results evidence that the O₂ estimator presents a good accuracy versus experiments and offers significant opportunities for improving engine control and after-treatment devices management during transient operation.