We present an empirical template spectrum suitable for fitting and subtracting/studying the Fe II and Fe III emission lines in the rest frame ultraviolet spectra of quasars and active galactic nuclei, the first empirical ultraviolet iron template to cover the full range of 1250-3090 A. Iron emission is often a severe contaminant in optical-ultraviolet spectra of active galactic nuclei and quasars. Its presence complicates and limits the accuracy of measurements of both strong and weak emission lines and the continuum emission, affecting studies of line and continuum interrelations, the ionization structure, and elemental abundances in active galaxies and quasars. Despite the wealth of work on modeling the quasar Fe II emission and the need to account for this emission in observed quasar spectra, there is no ultraviolet template electronically available to aid this process. The iron template we present is based on Hubble Space Telescope spectra of the narrow line Seyfert 1 galaxy, I Zwicky 1 (I Zw 1, z=0.061). The intrinsic narrow lines (>~ 900 km s-1) of this source and its rich iron spectrum make the template particularly suitable for use with most active galactic nuclei and quasar spectra. The iron emission spectrum, the line identifications, and the measurements of absorption and emission lines are presented and compared with the work of Laor et al. Comments on each individual line feature and the line fitting are available in Appendix A. The methods used to develop and apply the template are also described. We illustrate the application of the derived Fe II and Fe III templates by fitting and subtracting iron emission from the spectra of four high-redshift quasars and of the nearby quasar, 3C 273, confirming their general applicability to active galaxies despite the somewhat unusual properties of I Zw 1. We briefly discuss the small discrepancies between the observed iron emission of these quasars and the ultraviolet template, and compare the template with previously published ones. We discuss the advantages and limitations of the UV Fe II and Fe III templates and of the template fitting method. We conclude that the templates work sufficiently well to be a valuable and important tool for eliminating and studying the iron emission in active galaxies, at least until accurate theoretical iron emission models are developed. The Si IV+O IV] lambda1400 feature in I Zw 1 is clearly strong relative to C IV lambda1549, and C IV and C III] lambda1909 are both relatively weak. This may partially be due to the higher densities and lower ionization parameter prevailing in narrow line Seyfert 1 galaxies and to the big blue bump shifting toward lower energies in more luminous Seyferts, such as I Zw 1. In I Zw 1 the narrow line width reveals that C III] is heavily blended with Si III] lambda1892, Al III lambdalambda1854,1863, and Fe III transitions. This suggests that the C III] line strength and width may be overestimated in many quasar line studies where the lines are broader and deblending is not possible. This affects density estimates of the broad line region. Photoionization modeling, including all these line features, and subsequent fitting to the spectra are required to estimate the true C III] strength. We also argue, based on earlier work, that (strong) iron emission may be connected with high densities and associated with outflows. Based on observations made with the NASA/ESA Hubble Space Telescope, obtained from the data archive at the Space Telescope Science Institute. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under the NASA contract NAS 5-26555.