Background: Interventions to improve driving ability after stroke, including driving simulation and retraining visual skills, have limited evaluation of their effectiveness to guide policy and practice. Objectives: To determine whether any intervention, with the specific aim of maximising driving skills, improves the driving performance of people after stroke. Search methods: We searched the Cochrane Stroke Group Trials register (August 2013), the Cochrane Central Register of Controlled Trials (The Cochrane Library 2012, Issue 3), MEDLINE (1950 to October 2013), EMBASE (1980 to October 2013), and six additional databases. To identify further published, unpublished and ongoing trials, we handsearched relevant journals and conference proceedings, searched trials and research registers, checked reference lists and contacted key researchers in the area. Selection criteria: Randomised controlled trials (RCTs), quasi-randomised trials and cluster studies of rehabilitation interventions, with the specific aim of maximising driving skills or with an outcome of assessing driving skills in adults after stroke. The primary outcome of interest was the performance in an on-road assessment after training. Secondary outcomes included assessments of vision, cognition and driving behaviour. Data collection and analysis: Two review authors independently selected trials based on pre-defined inclusion criteria, extracted the data and assessed risk of bias. A third review author moderated disagreements as required. The review authors contacted all investigators to obtain missing information. Main results: We included four trials involving 245 participants in the review. Study sample sizes were generally small, and interventions, controls and outcome measures varied, and thus it was inappropriate to pool studies. Included studies were at a low risk of bias for the majority of domains, with a high/unclear risk of bias identified in the areas of: performance (participants not blinded to allocation), and attrition (incomplete outcome data due to withdrawal) bias. Intervention approaches included the contextual approach of driving simulation and underlying skill development approach, including the retraining of speed of visual processing and visual motor skills. The studies were conducted with people who were relatively young and the timing after stroke was varied. Primary outcome: there was no clear evidence of improved on-road scores immediately after training in any of the four studies, or at six months (mean difference 15 points on the Test Ride for Investigating Practical Fitness to Drive - Belgian version, 95% confidence intervals (CI) 4.56 to 34.56, P value = 0.15, one study, 83 participants). Secondary outcomes: road sign recognition was better in people who underwent training compared with control (mean difference 1.69 points on the Road Sign Recognition Task of the Stroke Driver Screening Assessment, 95% CI 0.51 to 2.87, P value = 0.007, one study, 73 participants). Significant findings were in favour of a simulator-based driving rehabilitation programme (based on one study with 73 participants) but these results should be interpreted with caution as they were based on a single study. Adverse effects were not reported. There was insufficient evidence to draw conclusions on the effects on vision, other measures of cognition, motor and functional activities, and driving behaviour with the intervention. Authors' conclusions: There was insufficient evidence to reach conclusions about the use of rehabilitation to improve on-road driving skills after stroke. We found limited evidence that the use of a driving simulator may be beneficial in improving visuocognitive abilities, such as road sign recognition that are related to driving. Moreover, we were unable to find any RCTs that evaluated on-road driving lessons as an intervention. At present, it is unclear which impairments that influence driving ability after stroke are amenable to rehabilitation, and whether the contextual or remedial approaches, or a combination of both, are more efficacious.