High densities of cyanobacteria in aquatic ecosystems can cause impacts to ecosystem services because they serve as a poor-quality food resource, produce toxins and can indirectly cause a variety of other negative impacts to water quality. There are many hypotheses about the potential environmental drivers of variation in cyanobacterial abundance and toxicity, but these hypotheses have rarely been considered in combination and rarely been examined in large river ecosystems. Here we use monthly data from backwater habitats of the Upper Mississippi River (UMR) to evaluate associations between environmental conditions and cyanobacterial abundance and toxicity (microcystin and anatoxin-a) that would be expected based on several hypotheses. Backwaters in the Mississippi River vary in flushing rate, temperature, turbidity, nutrient availability, water depth and vegetative cover. We find support for hypotheses that suggest physical conditions in backwaters (flushing rate, temperature, turbidity, rooted vegetation cover and water depth) and nutrient availability influence cyanobacterial abundance and toxicity. We then used structural equation modeling to incorporate several hypotheses into a causal modeling framework, which indicated that backwater connectivity (flushing) strongly influences cyanobacterial abundance via the regulation of water temperature, and that nutrient availability strongly influences the presence of microcystin concentrations above our detection limit. Our data suggest that management of backwater connectivity could influence cyanobacterial abundance and toxicity in UMR backwaters. Reconnecting backwaters (via alteration of levees) could serve as a local adaptation to minimize the effects of climate change and excessive nutrient loading.