Background Attentional bias (AB) to threat is thought to play a key role in the development and maintenance of posttraumatic stress symptomatology (PTS). Empirical evidence though is inconsistent. Some studies report associations between AB towards, threat and PTS; other studies report associations between AB away from threat and PTS; yet other studies fail to find any association. We propose that prospective–longitudinal study of AB as a dynamic process, expressed from moment to moment in time, may help to understand these mixed findings and the role of AB in PTS. Methods We tested cross‐sectional and prospective–longitudinal associations between AB and PTS among German soldiers from pre‐ to post‐deployment in Afghanistan (n = 144). AB to threat and positive emotion stimuli (angry/happy faces) was measured using the dot‐probe task. PTS was assessed by the PTSD Checklist. The number of traumatic experiences was assessed using CIDI‐traumatic experience lists for military. Results We found that AB dynamics (i.e., towards, away, temporal variability) at pre‐ and post‐deployment, with respect to angry and happy faces, predicted higher levels of PTS after deployment as a function of number of intermediate traumatic experiences. Traditional aggregated mean bias scores did not similarly prospectively predict PTS post deployment. Conclusions Findings indicate that AB to emotionally arousing stimuli may play an important function in the development and maintenance of PTS. We argue that mixed and null findings appear to be due to failure to model the within‐subject temporal variability in AB expression. Theoretical, empirical, and clinical implications of these findings are discussed.