Habitat use by waterfowl (breeding pairs and broods) of several Utah marshes was studied during 1978 and 1979. A number of ponds found on 5 different marshes located near Great Salt Lake were studied. Perimeter, area, depth, and emergent vegetative type were measured on each pond. Waterfowl use of the ponds was determined using a fixed point observation system. Multiple regression was used to analyze the relationships between the measured habitat variables and waterfowl use. Area of the pond was the variable most strongly related to both pair and brood use. Perimeter was highly correlated to area, but a calculated edge index reduced the confounding between the two variables. The edge index was positively related to pair use by 3 species of waterfowl and to total brood use, but it explained relatively little of the total variation in waterfowl use. Pond depth had little effect on duck use, but the type of emergent edge often had a significant effect on both pair and brood use. Waterfowl use of the ponds differed slightly between marshes. The multiple regressions explained from 53 to 79 percent of the variation in pair use by species. From 31 to 77 percent of the variation in brood use by group or species was explained. The 1978 data were used to validate the regression models developed with the 1979 data.