The effect of cannabis and cocaine use on non-fatal traffic injuries and use of motorcycle helmets or car seatbelts was assessed in a nationwide sample of 17 484 car or motorcycle drivers surveyed in 2005 in Spain. Logistic regression was used to adjust for distance driven and potential confounders. Cocaine use ≥1 day/week and cannabis use >4 days/week were associated with more traffic injuries. A positive dose–response relationship was found between frequency of cocaine use and lack of consistent use of protective devices. Interventions to avoid driving under the influence of drugs and to increase use of protective devices among drug users are needed.