The diversity and resultant habitat value of wetland plant communities in the Laurentian Great Lakes are dependent on water-level fluctuations of varying frequency and amplitude. Conceptual models have described the response of vegetation to alternating high and low lake levels, but few quantitative studies have documented the changes that occur. In response to recent concerns over shoreline management activities during an ongoing period of low lake levels in lakes Superior, Michigan, and Huron that began in 1999, we analyzed a quantitative data set from Saginaw Bay of Lake Huron collected from 1988 to 1993 during a previous lake-level decline to provide the needed information on vegetation responses. Transects were established that followed topographic contours with water-level histories that differed across a six-year period, ranging from barely flooded to dewatered for varying numbers of years to never dewatered. Percent cover data from randomly placed quadrats along those transects were analyzed to assess floristic changes over time, document development of distinct plant assemblages, and relate the results to lake-level changes. Ordinations showed that plant assemblages sorted out by transects that reflect differing water-level histories. Distinction of assemblages was maintained for at least three years, although the composition and positioning of those assemblages changed as lake levels changed. We present a model that uses orthogonal axes to plot transects by years out of water against distance above water and sorted those transects in a manner that matched ordination results. The model suggests that vegetation response following dewatering is dependent on both position along the water level/soil moisture gradient and length of time since dewatering. This study provided quantitative evidence that lake-level fluctuations drive vegetative change in Great Lakes wetlands, and it may assist in making decisions regarding shoreline management in areas that historically supported wetlands.