The Multi-Item Localisation (MILO) task provides a powerful framework for examining prospective planning and inhibitory tagging through sequential, goal-directed search behaviours (Thornton & Horowitz, 2004). Typical MILO studies use overlearned symbolic stimuli such as numbers or letters. The current work adapted the MILO task to present emotional facial expression sequences, to test how expressive faces influence search organisation, planning efficiency, and inhibitory control. Across three online experiments (total N = 144), participants searched through graded emotional sequences comprising positive–neutral–negative or negative–neutral–positive expressions using real faces with real expressions (Experiment 1), real faces with morphed expressions (Experiment 2), or synthetic MetaHuman faces with morphed expressions (Experiment 3). Serial Reaction Time (SRT) patterns for Vanish and Remain trials always followed parallel trajectories, replicating previous work and indicating that inhibitory tagging is independent of stimulus complexity. Participants showed a clear switch cost at the neutral face, suggesting that emotional sequences were segmented into two categories rather than being processed as a continuous gradient. Synthetic faces yielded smoother SRT trajectories, faster completion times, and higher search fluency than real faces. These findings show that MILO principles extend robustly to expressive faces and highlight synthetic faces as a valuable tool for reaction-time paradigms requiring precise perceptual control. The adapted MILO paradigm offers a novel behavioural method for probing emotional processing, perceptual fluency, and sequential organisation in complex visual domains.