The purpose of the current study was to determine the validity and reliability of an inertial sensor for assessing speed specific to athletes competing in the wheelchair court sports (basketball, rugby, and tennis). A wireless inertial sensor was attached to the axle of a sports wheelchair. Over two separate sessions, the sensor was tested across a range of treadmill speeds reflective of the court sports (1.0 to 6.0 m/s). At each test speed, ten 10-second trials were recorded and were compared with the treadmill (criterion). A further session explored the dynamic validity and reliability of the sensor during a sprinting task on a wheelchair ergometer compared with high-speed video (criterion). During session one, the sensor marginally overestimated speed, whereas during session two these speeds were underestimated slightly. However, systematic bias and absolute random errors never exceeded 0.058 m/s and 0.086 m/s, respectively, across both sessions. The sensor was also shown to be a reliable device with coefficients of variation (% CV) never exceeding 0.9 at any speed. During maximal sprinting, the sensor also provided a valid representation of the peak speeds reached (1.6% CV). Slight random errors in timing led to larger random errors in the detection of deceleration values. The results of this investigation have demonstrated that an inertial sensor developed for sports wheelchair applications provided a valid and reliable assessment of the speeds typically experienced by wheelchair athletes. As such, this device will be a valuable monitoring tool for assessing aspects of linear wheelchair performance.