Abstract: This paper presented a numerical methodology for recognition and extraction of artificial micro-geomorphology in the Jianghan Plain. In the different spatial scales of geomorphologic types, micro-geomorphology is the smallest spatial pattern and thus is more easily influenced by human activities. In this sense, artificial micro-geomorphology would show more potential in understanding the interactions between natural environment and human activities. Due to the continuous land utilization and rearrangement from historical period in the Jianghan Plain, a kind of artificial micro-geomorphology characterized by enclosed embankments on the low-lying land was formed. For decades, it played a very important role in the spatial distributions of some disasters such as flood and waterlogging. Compared with the automated extraction of drainage network and catchments boundaries from nature valleys, automated recognition of artificial micro-geomorphology skeleton lines has distinctive difficulties because of its tiny size, greater susceptibility to data error, and the different spatial structures with nature geomorphology. Despite of the progress in the automatic extraction of terrain skeleton lines, there is no auto-extracting method of artificial micro-geomorphology availablility yet, and it still dependes on the manual way to extract the skeleton lines of artificial geomorphology from the remotely sensed images and topographic maps. In this paper, a digital method for extracting the skeleton line of artificial micro-geomorphology was developed through the combination of the local topography analysis and the overland flow simulation. According to the specific artificial geomorphology features in the Jianghan Plain, a method was brought forward to identify the potential positions of the topographic skeleton lines, which is based on the evaluation of difference between the local elevations and the average elevations within moving windows. Then, an algorithm of watershed-merging based on flow simulation was developed to extract the potential skeleton lines of enclosed embankments. The algorithm can reduce the influence of the spurious pits in DEM, and generate the possible skeleton lines in a continuous way. Afterwards, based on the calculation results from the local topography analysis and the flow simulation, a new method was proposed to extract the skeleton lines of artificial micro-geomorphology by using the GIS techniques of spatial analysis and geo-statistics. The potential skeleton lines extracted from watershed-merging process were divided into segments by intersection points, and then transformed to raster format. Afterwards, raster segments of skeleton lines belonging to the proper skeleton lines of artificial micro-geomorphology were judged and fixed by using the position information calculated from local topography analysis according to their superposition degree. For the disconnected places of the extracted skeleton lines, a method was designed which could automatically extend and join the arcs on direction of the catchments borderlines through a kind of path determination algorithm. At last, the complete and closed borderlines of all the artificial geomorphologic patterns were formed through the above steps. The result shows that the aforementioned methodology is an effective attempt for the digital extraction of artificial micro-geomorphology structures, and would be helpful for the estimation of the interactions among spatial patterns of disaster, human activities and natural environment.