Phylogenetics is undergoing a revolution as large-scale molecular datasets reveal unexpected but repeatable rearrangements of clades that were previously thought to be disparate lineages. One of the most unusual clades of fishes that has been found using large-scale molecular datasets is an expanded Syngnathiformes including traditional long-snouted syngnathiform lineages (Aulostomidae, Centriscidae, Fistulariidae, Solenostomidae, Syngnathidae), as well as a diverse set of largely benthic-associated fishes (Callionymoidei, Dactylopteridae, Mullidae, Pegasidae) that were previously dispersed across three orders. The monophyly of this surprising clade of fishes has been upheld by recent studies utilizing both nuclear and mitogenomic data, but the relationships among major lineages within Syngnathiformes remain ambiguous; previous analyses have inconsistent topologies and are plagued by low support at deep divergences between the major lineages. In this study, we use a dataset of ultraconserved elements (UCEs) to conduct the first phylogenomic study of Syngnathiformes. UCEs have been effective markers for resolving deep phylogenetic relationships in fishes and, combined with increased taxon sampling, we expected UCEs to resolve problematic syngnathiform relationships. Overall, UCEs were effective at resolving relationships within Syngnathiformes at a range of evolutionary timescales. We find consistent support for the monophyly of traditional long-snouted syngnathiform lineages (Aulostomidae, Centriscidae, Fistulariidae, Solenostomidae, Syngnathidae), which better agrees with morphological hypotheses than previously published topologies from molecular data. This result was supported by all Bayesian and maximum likelihood analyses, was robust to differences in matrix completeness and potential sources of bias, and was highly supported in coalescent-based analyses in ASTRAL when matrices were filtered to contain the most phylogenetically informative loci. While Bayesian and maximum likelihood analyses found support for a benthic-associated clade (Callionymidae, Dactylopteridae, Mullidae, and Pegasidae) as sister to the long-snouted clade, this result was not replicated in the ASTRAL analyses. The base of our phylogeny is characterized by short internodes separating major syngnathiform lineages and is consistent with the hypothesis of an ancient rapid radiation at the base of Syngnathiformes. Syngnathiformes therefore present an exciting opportunity to study patterns of morphological variation and functional innovation arising from rapid but ancient radiation.