Stringent vehicles emissions regulations are nowadays leading to an increasing complexity in automotive products, in order to comply to the new policies about pollutants and fuel economy while preserving performances and driving comfort. Vehicles hybridization is also taking hold as an effective way to reach these new standards. In this context, simulation-based virtual calibration processes play a key role in time and costs reduction, allowing faster calibration and validation with respect to traditional test bench activities, also resulting in shorter time-to-market. This work focuses on the virtual calibration of the Engine Management System (EMS) torque estimation algorithm for a twin-turbocharged McLaren Spark-Ignition (SI) engine.