In many countries strict legal requirements for obtaining a driver’s license are in effect for visual acuity and visual field.We studied the relationship between these characteristics and driving safety and driving proficiency in an on-the-road test of practical fitness to drive in subjects with visual disorders, including many subjects scoring below current criteria. We further studied how far the relationship between the on-the-road test and visual measures improved if compensatory eye movements and visual attention were included in the criteria. Lastly, we studied the effects of training compensatory viewing strategiesformed the on-the-road test before and after training. Training consisted of laboratory and mobility training, including driving instruction. Visual function assessment included acuity, visual field, contrast sensitivity, visual attention, compensatory viewing efficiency, and visuospatial tests. In one study an advanced driving simulator was used besides the on-the-road assessment. Two models were compared to predict the on-the-road score. Results: 13–62% of the subjects passed the on-the-road test before training. After training, an additional 15–45% passed. The power of both models to predict the on-the-road score rose to about 45% by adding viewing behavior in the driving simulator. Discussion: A considerable percentage of the subjects, legally not allowed to drive, passed the on-the-road test. Sensitivity and specificity of vision tests and driving simulator tests are still too low to decide upon unfitness to drive. Training of compensatory viewing improved the performance in the on-the-road test.