Pseudo-nitzschia spp. produces a neurotoxin, domoic acid (DA), which can cause illness in humans, commonly known as amnesic shellfish poisoning, mass-mortality of marine animals, and closure of commercial shellfish fisheries during bloom events. Understanding and forecasting these harmful algal blooms is a primary management goal however, accurately predicting the onset and severity of Pseudo-nitzschia bloom events remains difficult, in part because the underlying drivers of bloom formation have not been fully resolved. Recent work suggests that the genetic composition of a Pseudo-nitzschia bloom may be a better predictor of toxicity than prevailing environmental conditions. Here, we test the ability of reduced representation sequencing (2b-RAD) to delineate inter and intra-species level genetic variation of Pseudo-nitzschia. We generated 2bRAD libraries from pure cultures of Pseudo-nitzschia spp. and mock community mixes. We also generated sequencing libraries from longitudinal field samples from the San Pedro Shelf. This submission contains all raw fastq data for the pure culture reference libraries and mock community mixes.