The Smart Fire Safety for Ships (SFSS) project focuses on improving fire situational awareness in the compartments of  naval ships. By enabling faster fire detection and early assessment of fire severity, the response time to incidents can be significantly reduced. A key objective is to estimate the size of the fire and its potential spread as quickly as possible, allowing rapid prioritization between multiple simultaneous incidents. This supports the timely deployment of appropriate resources and ultimately increases the survivability of the ship. This research performs fire detection through video in the visible spectrum and applies machine learning (ML) to predict the Heat Release Rate (HRR). The predicted HRR serves as input for a second ML-model, trained on numerical simulations, which forecasts the temperature distribution in the compartment of fire origin and in adjacent compartments. The overall concept will be validated through a proof-of-concept by conducting experiments in the Damage Control Centre of the Belgian Navy.