Innovations for terrestrial transportation technologies, e.g., cars, aircraft, and so on, have driven historical industries so far, and a similar breakthrough is now occurring in space owing to the successful development of electric propulsion devices such as gridded ion and Hall effect thrusters, where solar power is converted into the momentum of the propellant via acceleration of the ionized gases, resulting in a high specific impulse. A magnetic nozzle (MN) radiofrequency (rf) plasma thruster consisting of a low-pressure rf plasma source and a MN is an attractive candidate for a high-power electric propulsion device for spacecraft, as it will provide a long lifetime operation at a high-power level due to the absence of an electrode exposed to the plasma and a high thrust density. The high-density plasma produced in the source is transported along the magnetic field lines toward the open-source exit and the plasma is then spontaneously accelerated in the MN. By ejecting the plasma flow from the system, the reaction forces are exerted to the thruster structure including the source and the MN, and the spacecraft is resultantly propelled. The thruster will open the next door for space technologies, while the performance of the MN rf plasma thruster has been lower than those of the mature electric propulsion devices due to the energy loss to the physical walls. Here the thruster efficiency of about 30%, being the highest to date in this type of thruster, is successfully obtained in the MN rf plasma thruster by locating a cusp magnetic field inside the source, which acts as a virtual magnetic wall isolating the plasma from the source wall. The increase in the thrust by the cusp can be explained by considering the reductions of the loss area and the plasma volume in a thrust analysis combining a global source model and a one-dimensional MN model.