Anxiety is an emotional state occurring in threatening circumstances that affect perceptual and motor behavior. Entropy is a metric derived from information theory that indicates the randomness of a system and it has been used to quantify the drivers’ visual scanning complexity. Previous studies have been shown that visual entropy changes with increasing driver age and task processing demands. The aim of this study was to investigate the effects of anxiety on drivers’ visual scanning behavior. Sixteen experienced drivers (26.38 ± 2.80 years old) drove three minutes in a simulator, aiming to maintain the car speed between 100 and 120 km/h and to avoid collisions, under low- and high-anxiety conditions. Anxiety was manipulated using competition between participants, the presence of an evaluator, external video camera, and traffic noise. The frequency of fixations transitions between AOIs (lane, speedometer, and mirrors) was characterized by calculating first-order transition matrices. In addition, transitions probability were calculated using 1st order Markov process and adjusted into the Entropy equation. Drivers self-reported more anxiety score on STAI-S (13.52 ± 0.53 pts) and had higher mean Heart Rate (91.41 ± 2.62 bpm) during high-anxiety than low-anxiety condition (STAI-S = 09.37 ± 0.53 pts; HR = 79.43 ± 2.04 bpm), confirming the experimental manipulation. Student’s t-test revealed that, under high-anxiety condition, drivers made more fixations transitions between lane and speedometer AOIs (p = .001) and decreased the transitions between lane and mirrors (p = .003) compared to the low-anxiety. The frequency of transitions between speedometer and mirrors was unchanged with anxiety. Student’s t-test indicated that under high-anxiety condition the entropy of the fixations transitions between AOIs increased (p = .049), indicating that drivers’ visual scanning became more random under pressure. A higher visual scanning randomness implies into a poorer acquisition of information as well as a higher effort spent. Furthermore, the visual scanning is controlled by the goal-directed attentional system in order to direct attention towards task-relevant information. According to the Attentional Control Theory (ACT), anxiety reduces the goal-directed and intensify the stimulus-driven attention. Although anxiety evoked a restricted visual scanning between lane and speedometer, the entropy results, therefore, suggest an impaired top-down control on drivers’ attention under pressure. Acknowledgments FAPESP process n. 2015/10851-3.