Hillslope type Debris Flows (HDFs) often occur at the heritage in mountain areas, causing the serious impact to residents. However, the topography of HDF and shallow landslide is quite similar thus difficult to distinguish. This study proposal a HDF topographic recognition model based on the FISHER discrimination methodusing40 HDFs and 40 landslides induced by Typhoon Morakot in Kaoping River, Taiwan.The topographic recognition model consists of 7 variables, including effective watershed area index, average channel gradient of the transport segment, form factor rate, depression ratio of the transport segment, average width of the initiation segment, and average gradient of watershed, and its classification rate reached 90%. It is expected that the recognition mode will improve the slope disaster prevention of HDF type.