Every year many people all over the world are killed and severely injured in road traffic accidents. Even though driving too fast is a behaviour well known to contribute to both the number and the outcome of these accidents, drivers are still speeding. The general aim of this thesis, and its five empirical studies, is therefore to further the knowledge about drivers speeding behaviour by using the theory of planned behaviour and the model underpinning the driver behaviour questionnaire as frames of reference. The behavioural data used is obtained from field trials with intelligent speed adaptation and the speed reducing potential of this system is also examined. The results show that attitude towards exceeding the speed limits, subjective norm, perceived behavioural control and moral norm from the theory of planned behaviour, but also violations and inattention errors from the model underpinning the driver behaviour questionnaire, can be used to predict drivers’ everyday speeding behaviour. These two models can also be combined in order to gain further knowledge about the causes of speeding. Identification of drivers’ beliefs about exceeding the speed limits gives further insight into the underlying cognitive foundation of their attitude, subjective norm and perceived behavioural control. This provides valuable information for future design of speed reducing measures. Regarding intelligent speed adaptation, the results show that the ISA speed-warning device greatly reduces the amount of time drivers spend above the speed limits, and to some extent also reduces their mean speeds, but that this effect decreases with time. Although the drivers are not totally satisfied with the experience of the ISA speed-warning device, they like the idea and can see its usefulness. As the device tested is a first generation ISA speed-warning device, further research has the potential to greatly improve the system.