We present phase-resolved low-resolution infrared spectra of the polar EF Eridani obtained over a period of 2 yr with SpeX on the IRTF. The spectra, covering the wavelength range 0.8 μ m <= λ <= 2.4 μ m , are dominated by cyclotron emission at all phases. We use a "constant lambda" prescription to attempt to model the changing cyclotron features seen in the spectra. A single cyclotron emission component with Bsimeq 12.6 MG and a plasma temperature of kTsimeq 5.0 keV does a reasonable job in matching the features seen in the H and K bands, but fails to completely reproduce the morphology shortward of 1.6 μm. We find that a two-component model, where both components have similar properties but their contributions differ with viewing geometry, provides an excellent fit to the data. We discuss the implications of our models and compare them with previously published results. In addition, we show that a cyclotron model with similar properties to those used for modeling the infrared spectra, but with a field strength of B = 115 MG, can explain the GALEX observations of EF Eri.