This thesis documents the linear mixed model developed for vehicle speed along two-lane two-way rural horizontal curves in the outside lane. Speed data at each curve was collected at four points along the curve including the midpoint of the curve for a minimum of 48 hours during weekdays. Vehicle speed was analyzed separately for day and night conditions. The horizontal curves were categorized into different groups using different methods using side friction demand, radius and pavement edgeline marking retroreflectivity. In the speed prediction model, radius, superelevation at the midpoint of the curve, deflection angle, posted speed limit and pavement edgeline marking retroreflectivity were used to predict the vehicle speed at the midpoint of the horizontal curve. The regression analysis indicates that all of these variables are statistically significant in predicting the vehicle speed at the midpoint of horizontal curves with a 95 percent confidence interval. The linear model determined that the vehicle speed has a positive relation with the radius of the curve, superelevation and posted speed limit but has a negative relation with the deflection angle and pavement edgeline marking retroreflectivity. Curves were categorized based on side friction demand or radius and retroreflectivity of pavement edgeline marking. ANOVA was used to compare the day and night time speed. The comparisons reveal that vehicle speed at the horizontal curves decreases as the side friction demand value of the curves increases. Another finding of this research was that even though the posted speed limit is incorporated into the calculation of side friction demand, it may be necessary to analyze the impact of posted speed limit on vehicle speed for both daytime and nighttime. Previous literature determined that drivers may drive at an unsafe speed during nighttime at high levels of retroreflectivity. The results of this study could not confirm this statement as data from this study suggests that for curves with pavement edgeline marking retroreflectivity greater than 90 mcd/m2/lx, the effects of retroreflectivity on speed was determined to be minimal. This is based on the finding that the daytime and nighttime speeda were basically the same as the daytime and nighttime speed difference was both statistically and practically insignificant.