Existing thrust measurement systems (TMSs) at NASA Stennis Space Center use strain gauges and flux plates to measure forces produced by a test article. Alignment and calibration can take two weeks or more every time a piece of hardware or test article is changed. Cross axis loading is also problematic because it is impossible to perfectly align the flex plates and strain gauges in the thrust direction. In response to these problems, a magnetically levitated thrust measurement system has been proposed and a 300lb capacity demonstrator has been designed and built. In this design, the magnetic bearings work concurrently as support bearings and force measurement devices. The demonstrator consists of a floating frame that is completely levitated within a fixed frame by four support bearings carrying loads in the x- and y-direction and seven thrust bearings carrying loads in the z- or thrust direction. Joe Imlach of Imlach Consulting Engineering designed the demonstrator and magnetic bearing components, while Virginia Tech's role has been the application of the multipoint calibration technique including code development, the implementation of a 128-channel data acquisition system, and the overall test verification of the TMS demonstrator.A turnbuckle assembly and magnetostrictive actuator are used in series with a conventional load cell for static and dynamic testing, respectively. Both current based and flux based force equations were used to measure the reaction forces at the bearings. The static results using the current based equations including the current based fringing equations resulted in accuracies of 93% of full load, while the static results using the flux based equations including the flux based fringing equations resulted in accuracies of 99.5% of full load. These accuracies can be compared to accuracies of 83-90% seen in previous work using magnetic bearings to measure forces by monitoring currents and to accuracies of about 99% in previous work using magnetic bearings to measure forces by monitoring fluxes. All of the improved accuracies were made possible through the implementation of a calibration technique known as the multipoint method and the implementation of a gap dependent fringing correction factor developed by Joe Imlach. The demonstrator was not outfitted with accelerometers so the inertia of the floating frame could not be accounted for, limiting the scope of dynamic testing. However, the tests confirmed the ability of the demonstrator to measure dynamic loads in general. / Master of Science