Nighttime crashes are over-represented on the US highway system. Roadway lighting, providing additional visibility by supplementing vehicle headlights, has been identified as an effective countermeasure to improve nighttime safety. However, the effect of street lighting illuminance in reducing the injury severity of nighttime crashes on roadway segments is not well-documented. This study aimed to investigate the effect of street lighting illuminance, rather than the presence of street lighting, on nighttime crash severity on roadway segments. Illuminance data were collected in the Tampa Bay area in Florida from 2012–2014 using the Advanced Lighting Measurement System and four years of crash data (2011–2014). A random parameter ordered probit model was developed based on the collected data for addressing the unobserved heterogeneity issue in the sample. The major conclusions include the following: (1) a medium illuminance level (0.4–0.8 fc) compared to a low illuminance level (< 0.4 fc) can significantly reduce the injury severity of nighttime crashes; (2) a high illuminance level (> 0.8 fc) has no significant influence on the injury severity of nighttime crashes compared to a medium illuminance level, but its effect is random (63.7% increase in injury severity and 23.7% decrease); (3) the involvement of vulnerable road users (pedestrians, bicyclists, motorcyclists) and aggressive driving are the first and second most significant factors contributing to severe injury (fatal or incapacitating) in nighttime crashes; and (4) other significant factors include injured party gender, driver age (at fault), crash type, roadway speed limit, lane configuration, pavement friction, etc.