A validation study of computational fluid dynamics for supersonic retropropulsion was conducted using three Navier–Stokes flow solvers. The study compared results from the computational-fluid-dynamics codes to each other and to wind-tunnel test data obtained in the NASA Ames Research Center Unitary Plan Wind Tunnel. Comparisons include surface pressure coefficient as well as unsteady plume effects and cover a range of Mach numbers, levels of thrust, and angles of orientation for zero-, one-, three-, and four-nozzle configurations. Flow-structure behavior changed with thrust and angle of orientation for all nozzle configurations. In general, the solvers compared best with the test data for the steadier cases of the one-nozzle and high-thrust three-nozzle configurations. Deviation in surface pressure was noted for the more unsteady cases and near transitions in behavioral modes. Strengths and weaknesses of the solvers are identified, and possible error sources are discussed.