While speeding is considered as one of the most important road accident contributory factors, the objective of the present study is to investigate which driving performance parameters affect speeding. In order to achieve this objective a naturalistic driving experiment took place where almost 5.000 trips of 116 drivers were recorded. Then 3 linear regression models were developed aiming to investigate which factors affect speeding which is defined as the percentage of the trip that the driver speed was exceeding the speed limit. More specifically, an overall model is developed for all road types, as well as separate models for urban and rural road type, respectively. Results reveal correlations of speeding percentage with specific driving behaviour and exposure metrics, namely the average speed, the number of harsh acceleration events, the eco-friendly driving performance as well as driving on weekends. The aforementioned analysis results constitute interesting findings and can be very important in policy makers and enforcement authorities in order to deal with the high risk factor of speeding.