Acoustic seabed classification is a useful tool for monitoring marine benthic habitats over broad-scales (>1 km2) and meso-scales (10 m2–1 km2). Its utility in this context was evaluated using two approaches: by describing natural changes in the temporal distribution of marine biotopes across the broad-scale (4 km2), and by attempting to detect specific experimentally-induced changes to kelp-dominated biotopes across the meso-scale (100 m2). For the first approach, acoustic backscatter mosaics were constructed using sidescan sonar and multibeam echosounder data collected from Church Bay (Rathlin Island, Northern Ireland) in 1999, 2008 and 2009. The mosaics were manually segmented into acoustic facies, which were ground-truthed using a drop-video camera. Biotopes were classified from the video by multivariate exploratory analysis and cross-tabulated with the acoustic facies, showing a positive correlation. These results were integrated with bathymetric data to map the distribution of seven unique biotopes in Church Bay. Kappa analysis showed the biotope distribution was highly similar between the biotope maps, possibly due to the stability of bedforms shaped by the tidal regime around Rathlin Island. The greatest biotope change in this approach was represented by seasonal and annual changes in the growth of the seagrass, Zostera marina. In the second approach, sidescan sonar data were collected before and after the removal of 100 m2 of kelp from three sites. Comparison of the data revealed no differences between the high-resolution backscatter imagery. It is concluded that acoustic seabed classification can be used to monitor change over broad- and meso-scales but not necessarily for all biotopes; its success depends on the type of acoustic system employed and the biological characteristics of the target biotope.