Decision making (DM) is essential and proven to be a natural and inherent part of the success of transport systems, particularly given the fast growth of autonomous systems in transport. It is critical but remains challenging to understand and predict DM performance in transport, because operators’ mental states have not been effectively considered in complex DM processes such as ship anti-collision operations. This paper proposes an advanced decision support methodology that pioneers the incorporation of objective neurophysiological and subjective data to analyse functional connectivity in the brain and predict DM performance in ship navigation. Experiments were conducted using a functional Near-Infrared Spectroscopy (fNIRS) technology to explore the functional connectivity of two groups (low workload and high workload) and predict their DM performance in a ship collision avoidance situation. It brings brain science into transport engineering and the results generate new contributions to the existing knowledge, including (1) the establishment of a methodology to detect different workload levels in safety–critical transport systems using psychophysiological measurement; (2) analysis of brain’s functional connectivity of different groups of decision makers (e.g., seafarers) with high and low workload tasks; (3) an advanced methodology to assess human reliability in complex scenarios and predict operational behaviours; (4) pioneering a human-centred approach to predict DM performance and demonstrate its feasibility in shipping. From a practical perspective, stakeholders can utilise the findings of this study to rationally evaluate human performance in transport system operations, aiding in operator qualification and certification processes. Furthermore, it is critical for adaptive automation regarding DM support in safety–critical systems.