Previous work indicated that automated camera traps may be useful in estimating predation risk among different microhabitats for mule deer (Odocoileus hemionus). We tested the prediction that the number of photographs taken by automated camera traps was inversely related to the amount of food left by deer in feeding boxes or giving up densities (GUDs). We positioned camera traps adjacent to standard mule deer feeding boxes placed in open and edge microhabitats of 3 forest types: Douglasfir (Pseudotsuga menziesii), juniper (Juniperus osteosperma), and mountain mahongany (Cercocarpus ledifolius). We compared number of photographs taken with daily GUDs for the boxes for the 2 microhabitats. We found that GUD data of our study coincided with the previous work of lower GUDs in open versus edge microhabitats for Douglas-fir (288.6±17.1 g vs. 389.6±19.4 g; P<0.001) and juniper (218.9±26.3 g vs. 251.9±29.6 g, P=0.027) but not for mountain mahogany (272.4±29.5 g vs. 287.0±32.3 g, P=0.414). We also found significantly more total photos/camera taken in open microhabitat versus edge microhabitat in Douglas-fir (16.2±2.2 vs. 7.4±1.5; P=0.004). More photos/camera were taken in the open versus edge in the juniper forest type (15.6±3.9 vs. 11.5±3.9), but the difference was not significant. There was no difference in photos/camera between the 2 microhabitats in the mountain mahogany (18.4 ± 3.4 vs. 19.3 ± 1.9). Total number of photos/day/box also was significantly related to the GUDs for the 3 forest types (P = 0.002–0.008). The amount of variability explained by regression equations (r2adjusted) ranged from 22% for Douglas-fir to 29% for juniper. We concluded that the total number of photographs taken does reflect the results of GUD analysis and that automated camera traps could be used to assess predation risk among different microhabitats.