The heterotrophic labyrinthulomycete protists have long been known to play an important role in the nutrient cycling of coastal seawater. Yet, their spatiotemporal abundance and diversity in polluted coastal waters remain poorly discussed, due in part to the paucity of a rapid detection method. To this end, we developed a qPCR detection method based on a newly designed primer pair targeting their 18S rRNA gene. Using this method, we studied the population dynamics of labyrinthulomycete protists in nutrient-rich (Shenzhen Bay) and low-nutrient (Daya) coastal habitats along the Pearl River Delta. We found a significantly (P < 0.05) higher abundance of Labyrinthulomycetes in the Shenzhen bay (average 3455 gene copies mL-1) than that in Daya Bay (average 378 gene copies mL-1). Their abundance gradient positively correlated (P < 0.05) with the levels of inorganic nitrogen and phosphates. Further characterization of the molecular diversity of these protists in Shenzhen Bay using different primer sets revealed the presence of several genera besides a large number of unclassified OTUs. Regardless of the primer biases, our results show significant (P < 0.05) spatiotemporal changes in the molecular abundance and diversity of these heterotrophic protists. Overall, this study provides a rapid molecular detection tool for Labyrinthulomycetes and expands our current understanding of their dynamics controlled by physicochemical gradients in coastal waters.