The Great Salt Lake (GSL) is a unique hypersaline system with an understudied phytoplankton assemblage supporting a productive open water ecosystem in the largest embayment of the lake, Gilbert Bay. Determination of phytoplankton by microscopy has practical limitations that can constrain the scope of a study, but DNA metabarcoding may improve upon this through higher taxonomic resolution and the capacity to generate a large volume of assemblage data in comparatively little time. To determine if metabarcoding could replicate microscopy and expand the assessment of GSL phytoplankton, a 23S SSU rRNA metabarcoding and microscopy survey of Gilbert Bay was conducted in 2017 and 2018. Assemblage composition and relative abundances from each method were compared, and spatial and temporal assemblage changes from metabarcoding data were investigated using non-metric multidimensional scaling. Metabarcoding differed from microscopy in multiple taxonomic assignments and relative abundances, with poor correlation for most categories. Diatoms were overrepresented by metabarcoding relative to microscopy, and chlorophytes underrepresented. However, metabarcoding revealed seasonal and spatial patterns in assemblage, detected seasonal patterns within phytoplankton sequences of very low abundance, and detected potential cryptic speciation within the lake’s dominant Dunaliella viridis. Phylogenetic analysis revealed greater phytoplankton diversity than observed before in GSL, but demonstrated the need to improve taxonomic assignment of the resulting sequences, particularly within the diatoms. The expansion of detectable diversity and isolation of DNA sequences that can be traced through time and analyzed against environmental variables make metabarcoding a potentially effective tool for parallel use with microscopy in future GSL research.