The parasitic sea lamprey (Petromyzon marinus) has caused extensive losses to commercial fish stocks of the upper Great Lakes of North America. Methods of controlling the sea lamprey include trapping, barriers to prevent migration, and use of a chemical lampricide (3-trifluoromethyl-4-nitrophenol) to kill the filter-feeding larvae. Concerns about the non-specificity of these methods have prompted continued development of species-specific methods to control lampreys outside their native range. In this study, we considered the utility of RNA interference to develop a sea lamprey-specific lampricide. Injection of six different short interfering, double-stranded RNAs (siRNAs) into lamprey embryos first confirmed that the siRNAs could reduce the targeted transcript levels by more than 50%. Two size classes of lamprey larvae were then fed the siRNAs complexed with liposomes, and three of the siRNAs (targeting elongation factor 1α, calmodulin, and α-actinin) reduced transcript levels 2.5, 3.6, and 5.0–fold, respectively, within the lamprey midsections. This is not only the first demonstration of RNAi in lampreys, but it is also the first example of delivery of siRNAs to a non-mammalian vertebrate through feeding formulations. One of the siRNA treatments also caused increased mortality of the larvae following a single feeding of siRNAs, which suggests that prolonged or multiple feedings of siRNAs could be used to kill filter-feeding larvae within streams, following development of a slow-release formulation. The genes targeted in this study are highly conserved across many species, and only serve as a proof-of-concept demonstration that siRNAs can be used in lampreys. Given that RNA interference is a sequence-specific phenomenon, it should be possible to design siRNAs that selectively target gene sequences that are unique to sea lampreys, and thus develop a technology to control these pests without adversely affecting non-target species.