Unmanned aerial vehicles (UAVs) are commonly utilized as cost-effective devices for data collection by capturing photos of target objects. UAV images have been used for many applications, such as civil engineering, transportation, architecture, surveying, and mapping. Although commercial UAV image data processing software is suitable for generating orthoimages and dense point clouds of surfaces, it still requires extensive labor to prepare the appropriate point cloud to create a digital elevation model (DEM). This study proposes a method to automatically create DEM from a point cloud generated from UAV images. The proposed method composes of three main steps: (1) Candidate ground points, (2) Ground points extraction, and (3) Creation of a DEM model. The proposed method was tested on three datasets, covering a total area of approximately 45 hectares from 200 images captured by DJI Phantom 4 drone. As a result, the DEMs are successfully created with a spatial resolution of 1.0 m.