The dynamics of phytoplankton community structure in two agricultural irrigation ponds located in Maryland, USA were evaluated. Stable spatiotemporal patterns and zones of consistently higher and consistently lower phytoplankton functional group concentrations were established for both ponds. Moderate and strong correlations were found between the spatial patterns of several water quality parameters and phytoplankton concentrations. Additionally, zones of consistently higher and lower concentrations were found for the cyanobacteria pigment, phycocyanin. Chlorophyll, colored dissolved organic matter, and turbidity were the most influential predictors for phycocyanin concentrations. The prediction of phytoplankton community structure from water quality measurements with the random forest machine learning algorithm was possible and easily measured physicochemical parameter models offered the best model performance. Results of this work indicate that in-situ water quality measurements may be a cost-effective and faster alternative to time-intensive microscopy analysis of phytoplankton, allowing for more efficient water quality monitoring.