Attentional Control Theory (ACT) suggests the negative performance impact of anxiety results from an initial reallocation of attentional resources, which decreases goal-directed control and increases stimulus-driven behaviour. Evidence from goal-directed aiming indicates that the negative effects of anxiety unfold near the end of movement during online control (Lawrence et al., 2013). We aimed to explore whether the anxiety effect in online control was also related to changes in offline planning. Participants aimed to a target under both low and high anxiety conditions. Following initial practice, participants were instructed to aim as fast-and-accurate as possible (low) or additionally received non-contingent feedback that previous responses were in the lower thirtieth percentile of the cohort (high). A manipulation-check (cognitive sub-scale of the Mental Readiness Form-3) indicated a significant increase in anxiety following the high condition. The performance outcomes (movement time, constant error) and movement kinematics (peak acceleration, peak velocity, peak deceleration, movement end) were assessed. There were no significant differences in any of the outcome-related measures. There was a shorter time to peak deceleration and longer displacement at peak velocity for the high compared to the low condition. To assess the error-reduction online control processes, we analysed the mean within-participant correlations between the displacement to and after specific kinematic landmarks. This analysis revealed a lower negative relation (less error-reducing) at peak deceleration for the high compared to the low condition. These findings indicate that high state anxiety initially alters the offline planning of action, which may result from an overload in attentional resources.