Over the past 20 years, fractal geometry has gained acceptance as a valid quantitative method for describing complex geomorphic patterns in nature. One goal of geomorphology is to relate geologic processes to the development and evolution of landforms. The processes (factors) that shape a landform might produce a characteristic geometry that can be interpreted through fractal analysis. The geometry of a shoreline is often cited as an example of a fractal pattern. Previous studies on shoreline geometries have used relatively small-scale maps and aerial photographs, and have not studied the changing geometry of a shoreline over time. This study, in contrast, examines the fractal geometry of six wave-dominated marsh shorelines in Rehoboth Bay, Delaware, at a relatively large scale, and investigates how these geometries change over time. The results help to understand the relation between geomorphic features and processes. An automated box-counting method was used to calculate fractal D values for each shoreline. D values ranged from 1.01 to 1.19, which are lower than previous shoreline studies. The greater D values are associated with sites that experienced the greatest wave erosion. Furthermore, the geometric complexity of these shorelines did not change over 3 years. In comparison, D values obtained from a tide-dominated marsh shoreline were greater (1.27), suggesting that the dominant formative process (e.g., tides vs. waves) may be distinguished through geometric analysis. However, to use fractal data as a comparable geomorphic tool, it is necessary to choose features of a similar size (scale) and from a similar geomorphic setting. In this way landform complexity (D) is likely due to either (1) the type of shaping process in different geographic settings, or (2) the magnitude of a single shaping process within a single geographic setting.