Inhibitory control deficits are considered a key pathogenic factor in anxiety disorders. To assess inhibitory control, the antisaccade task is a well-established measure that assesses antisaccade performance via latencies and error rates. The present study follows three aims: (1) to investigate inhibitory control via antisaccade latencies and errors in an antisaccade task, and their associations with multiple measures of fear in patients with spider phobia (SP) versus healthy controls (HC), (2) to investigate the modifiability of antisaccade performance via a fear-specific antisaccade training in patients with SP and HC, and (3) to explore associations between putative training-induced changes in antisaccade performance in SPs and changes in diverse measures of fear. Towards aim 1, we assess antisaccade latencies (primary outcome) and error rates (secondary outcome) in an emotional antisaccade task. Further, the baseline assessment includes assessments of psychophysiological, behavioral, and psychometric indices of fear in patients with SP and HCs. To address aim 2, we compare effects of a fear-specific antisaccade training with effects of a prosaccade training as a control condition. The primary and secondary outcomes are reassessed at a post-1-assessment in both SPs and HCs. Aim 3 employs a cross-over design and is piloted in patients with SP, only. Towards this aim, primary and secondary outcomes, as well as psychophysiological, behavioral, and psychometric measures of fear are reassessed at a post-2-assessment after the second training block. This study aims to better understand inhibitory control processes and their modifiability in spider phobia. If successful, antisaccade training may assist in the treatment of specific phobia by directly targeting the putative underlying inhibitory control deficits. This study has been preregistered with ISRCTN (ID: ISRCTN12918583) on 28th February 2022.