The National Compact Stellarator Experiment (NCSX) has been designed to accommodate a variety of heating systems, including ohmic heating, neutral beam injection, and radio-frequency (rf). Neutral beams will provide one of the primary heating methods for NCSX. In addition to plasma heating, neutral beams are also expected to provide a means for external control over the level of toroidal plasma rotation velocity and its profile. The experimental plan requires 3 MW of 50-keV balanced neutral beam tangential injection with pulse lengths of 500 ms for initial experiments, to be upgradeable to pulse lengths of 1.5 s. Subsequent upgrades will add 3 MW of neutral beam injection (NBI). This paper discusses the NCSX NBI requirements and design issues and shows how these are provided by the candidate PBX-M NBI system. In addition, estimations are given for beam heating efficiencies, scaling of heating efficiency with machine size and magnetic field level, parameter studies of the optimum beam injection tangency radius and toroidal injection location, and loss patterns of beam ions on the vacuum chamber wall to assist placement of wall armor and for minimizing the generation of impurities by the energetic beam ions. Finally, subsequent upgrades could add an additional 6 MW of rf heating by mode conversion ion Bernstein wave (MCIBW) heating, and if desired as possible future upgrades, the design also will accommodate high-harmonic fast-wave and electron cyclotron heating. The initial MCIBW heating technique and the design of the rf system lend themselves to current drive, so if current drive became desirable for any reason, only minor modifications to the heating system described here would be needed. The rf system will also be capable of localized ion heating (bulk or tail), and possibly IBW-generated sheared flows.