In recent years, the increasing competitiveness of motorcycle racing has made the adoption of advanced virtual tools essential to optimize engine performance and improve reliability. In motorsports, ongoing technological developments, including tightening of regulations in 2026, especially regarding the fuel-flow limit, have further increased the complexity of sustaining maximum performance. Within this framework, a predictive combustion model is essential to support engine development. It enables the early evaluation of different configurations and calibration strategies, while helping to identify effective solutions for maximizing performance under increasingly restrictive regulations. Therefore, this work aims to build a predictive combustion model and knock model for a high-performance four-cylinder spark-ignition engine, in collaboration with Ducati Motor Holding.