Global Navigation Satellite System (GNSS) receivers are now widely used for navigation and speed measurements. General public and companies which use GNSS rely not only on positional accuracy but also on speed accuracy. Manufacturers of GNSS receivers supply speed accuracy parameters in the relevant data sheets. However, little information is provided regarding specific conditions when the specified speed accuracy of GNSS receivers might be met. Also, little research was conducted to reveal the practical speed accuracy parameters of GNSS receivers in a variety of conditions focusing on challenging GNSS environments. Finally, no activities was conducted to understand if adding more constellations to GPS, for example, GLONASS provides any value in speed accuracy reporting. This research firstly aims to analyse and practically estimate the performance of high end, mid-range and low grade GNSS receivers for measuring speed in challenging environments and determine if practical differences in their speed accuracies are observed. Secondly, the research also aims to determine if adding GLONASS to GPS in the GNSS receiver’s computations provides any value in speed accuracy determination. Lastly, the research aims to derive a simple quality indicator which might be used to filter potentially unreliable GNSS speed records. After designing and calibrating the accurate non-GPS based test vehicle and analysing its uncertainty of measurements, a number of GNSS receivers were tested. Test results demonstrate that high end, mid-range and low grade GNSS receivers perform differently when measuring speed and caution should be exercised when relying on GNSS speed in challenging environments, specifically if such speed records are considered to be used in the court of law. It was also determined that for a specific receiver adding GLONASS does not improve the performance in speed accuracy. Finally, Horizontal Dilution of Precision (HDOP) parameter derived from GNSS receivers might be considered as a simple statistical quality indicator to assess whether specific speed records can be relied upon but cannot be used as an integrity indicator for individual speed records. The research recommends that every receiver shall be individually tested in a variety of environments to reveal its true errors in speed measurements.