We determined the feasibility of using unmanned aerial vehicle (UAV) video monitoring to predict intake of discrete food items of rangeland-raised Raramuri Criollo non-nursing beef cows. Thirty-five cows were released into a 405-m2 rectangular dry lot, either in pairs (pilot tests) or individually (experiment tests), that contained 12 food bowls arranged in an open semicircle and placed approximately 1 m apart. Four bowls containing either long alfalfa hay (AH, 200 g), long Sudangrass hay (SH, 200 g), or cottonseed cake (CC, 50 g) were alternated (CC, AH, SH) using the same sequence in all tests. Video footage of all arena tests was acquired with a three-dimensional Robotics Y6 Multi-copter fitted with a two-axis brushless gimbal and a GoPro Hero 3 Silver Digital Camera. Video files were processed to extract a total of 4893 two-second-interval still images that were viewed to determine cow feeding activity. Cows that were naïve to the sound of the UAV fed as frequently (P > 0.05) as their adapted counterparts during 12-min pilot tests. Significant positive correlations (r = 0.68-0.91; P < 0.05) between video-derived feeding frequency estimates and amount of AH, SH, and CC consumed per bowl were observed during the individual 4-min experiment tests. Our results suggest that UAV video monitoring could be a useful tool to monitor feeding behavior of rangeland cows.