Objective: This study examines the predictive ability of the CPT-II on total driving mistakes on a driving simulator in a normal adult population. Method: Participants included 108 males and 150 females with an average age of 28.69 years (SD = 11.55) and average education of 15.50 years (SD = 1.97). The data was derived from an ongoing de-identified database of mixed clinical and normal patients at an outpatient mental health clinic in South Florida. They were given the CPT-II and a driving simulation, which included an attention task requiring participants to notice shape changes in their periphery. Total driving mistakes were calculated from the sum of incorrect responses and missed responses on this task. Results: A linear regression analysis was used to analyze the predictive power of the CPT-II variables on total driving mistakes. The overall regression model in which total driving mistakes was predicted from the CPT-II variables was statistically significant F (14, 257) = 4.95 p < .001, and accounted for 22.4% of the variance in mistakes. The individual predictors in the model that were significant included Omission, Hit Rate, and Hit Rate Standard Error. Conclusion: Results demonstrate that higher levels of missed targets and inconsistent responding predicted greater total mistakes in attention. In addition, greater average speed for correct answers was associated with fewer mistakes, and vice versa. These results have important implications for making appropriate recommendations for safe driving as lowered ability to recognize and respond to stimuli, might lead to a higher likelihood of speeding and collisions.