Environmental DNA (eDNA) methods are increasingly applied in the marine environment to identify species and community structure. To establish widely applicable eDNA techniques for elasmobranchs, we used the Critically Endangered largetooth sawfish (Pristis pristis Linnaeus, 1758) as a model species for: (1) assessing eDNA particle size distribution; (2) assessing the efficiency of long-term preservation of water samples; and (3) comparing the efficiency and detection sensitivity of filtration and precipitation methods. Water samples (1 L) collected from a tank containing one largetooth sawfish specimen were sequentially filtered through five filter membranes of decreasing pore size (20, 10, 5, 1.2, and 0.45 μm). The proportion of sawfish eDNA retained within each size class was determined through quantitative real-time PCR (qPCR) using a species-specific TaqMan probe assay. A linear mixed-effects model (lme) showed that the 1.2 and 20 µm filters captured most of the eDNA particles present in the sampled water. Additionally, whole water samples (0.375 L) were preserved in Longmire's buffer, stored at tropical ambient temperatures (26.3°C ± 3.0 SD) and extracted at five time points: immediately, one, two, and three months after collection, as well as frozen and extracted three months later, to assess the preservation efficiency of Longmire's buffer via qPCR analysis. A linear mixed-effects model showed that samples maintained maximal eDNA yield for at least three months after collection at ambient storage. Lastly, when comparing the filtration and precipitation methods, filtration using 0.45 µm pore size was more sensitive to capture of largetooth sawfish eDNA than filtration with 20 µm filter or water precipitation. However, water precipitation was more efficient when accounting for volume of water processed. These results provide options for best capture and preservation of elasmobranch eDNA.