Marine biodiversity collections are critical resources for understanding biodiversity and evolutionary patterns. However, their taxonomic utility is limited due to challenges in morphological identification and cryptic speciation, particularly in sponges (Porifera). This study applied a novel target-enriched multilocus assay to a decades-old unidentified collection of Tethyidae from the Queensland Museum. Target-specific baits were designed based on Tethya genome data, enabling the sequencing of nuclear ultraconserved elements (UCEs) and mitochondrial genomes, even from degraded samples. In the study, 40 specimens were sequenced, and we found up to 4,440 UCEs per sample. The phylogenies were based on up to 2,788 loci and demonstrated overall high bootstrap support. Results include robust support for earlier findings on the non-monophyly of Tethya, with lineages linking to other genera (e.g., Stellitethya, Tethycometes), suggesting that Tethya is a paraphyletic group foundational to Tethyidae. Morphological subgroupings, previously hypothesized, were validated genetically, though several novel relationships emerged. This approach highlights the suitability of target-enriched multilocus assays for archival collections, enabling detailed biodiversity assessments using collection material initially not sampled for molecular work. Our Tethya case study underlines that this methodology provides a framework for broader applications in marine biodiversity conservation, facilitating the understanding and preservation of past and present biodiversity to guide future efforts.