Seagrasses are recognized as important plant communities in coastal estuaries and lagoons across both tropical and temperate climes; thus, large-scale seagrass die-off events worldwide are of general concern. In Florida Bay, at the southern terminus of the Florida peninsula, seagrass die-off events up to 4000 ha have been reported and smaller scale mortality events are noted annually. In the present study, we examined several hypothesized causative factors (high temperature, hypersalinity, sulfide toxicity) of seagrass ( Thalassia testudinum) mortality in Florida Bay. To test sulfide effects, in situ sulfide production was stimulated by applying a labile carbon source (glucose) to sulfate reducers in the sediment at five sites across the bay (northeastern, northcentral, and southwestern basins). During the one year study, high temperature (32–36 °C) and salinity (> 50 psu) were recorded in the bay associated with a regional drought. We also experienced major seagrass die-off events at two of our southwestern bay sites. These field conditions provided an excellent opportunity to closely examine cause–effect relationships among stressors and die-off events in the field, and verify results of our previous mesocosm experiments. Even though glucose amendments stimulated porewater sulfides in bay sediments (4–8 mmol L − 1 ), no significant differences in biomass, short shoot density or final growth rates were found between control and glucose plots. In addition, the highest growth rates and shoot densities were concomitant with maximum water column salinity (> 50 psu) and temperature (32–36 °C), when porewater sulfides were also in the millimolar range. Large-scale seagrass mortality events, encompassing ∼ 50% of the entire meadow at one site, occurred at southwestern bay sites when plants were down regulating (slower growth and shoot density), probably in response to shorter day length and lower temperature (30–34 to 23–26 °C) from October, 2004 to January, 2005. Sulfate reduction rates (SRR) were also 2-fold higher in the southwestern (214–488 nmol cm − 3 d − 1 ) versus northcentral and northeastern (97–240 nmol cm − 3 d − 1 ) bay sites, possibly limited by labile carbon, which we found to stimulate SRR 3-fold in northeastern and northcentral bay sites (461–708 nmol cm − 3 d − 1 ) and 4-fold at southwestern bay sites (1211–2036 nmol cm − 3 d − 1 ). Based on a synthesis of the field data reported herein, our mesocosm experiments to date, and contributions by others, we present a conceptual model of seagrass die-off in Florida Bay outlining a cascade of stressors, stimulated by P enrichment, which leads to high O 2 consumption in the system triggering a seagrass die-off event.