Speed information from GPS is increasingly used and provides an alternative to conventional methods such as wheel speed sensors. We investigate the possibility to use GPS derived speed as a reference when verifying laser and radar-based speed measuring devices used in traffic enforcement. We have set up a realistic test scenario where a GPS equipped vehicle was driven at three different speeds (40, 90 and 130 km/h) through a pre-defined measurement zone. An independent and traceable reference speed was calculated by accurately measuring the length of the measurement zone (approximately 15 metres), and the time it took to pass through it. The reference speed was compared to the average GPS speed for each passage. This comparisons show that the standard uncertainty of such GPS speed measurements is less than 0.05 km/h. Hence, GPS derived speed meets the accuracy requirements for verification of laser and radar based speed measuring devices.