<jats:p> The negative effect of reduced visibility on driver performance has been recognized as one of the main causes of motor vehicle crashes in fog. Although many studies have concentrated on driver behavior during foggy weather in a simulated environment, there is a lack of studies that have addressed the impact of fog on driver behavior and performance in naturalistic settings. This paper utilized the data from the SHRP2 Naturalistic Driving Study (NDS) database to understand driver behavior in general and speed selection in particular during clear and foggy weather conditions. In this study, a comparative preliminary analysis and an ordered logit model were developed to evaluate driver speed behavior in fog and clear weather conditions. Results from the preliminary analysis showed 10% and 3% reduction in speed because of near fog and distant fog, respectively. In addition, results from the speed selection model showed that the odds of reducing speed were 1.31 and 1.28 times higher for drivers traveling in near fog and distant fog, respectively, compared with drivers who were driving in clear weather conditions. However, there is an over-representation of young drivers in the SHRP2 NDS database, which was reflected in the dataset used in this study. Therefore, a more representative sample of age groups might provide different results. The results from this study could provide a better insight into driver speed selection during foggy weather conditions, which can be utilized to improve various safety strategies including variable speed limits. </jats:p>