We investigated fish assemblage structure in engineered (rip-rap) and natural habitats (log jams and mud banks) in the Kansas River USA to determine if natural structures had higher abundance and diversity of fishes at a local spatial scale. A total of 439 randomly selected sites were boat electrofished from May to August 2005 and 2006. Mean species diversity and richness were significantly higher in rip-rap than log jams and mud banks. Mean relative abundance (CPUE; number of fish collected per hour electrofishing) of six of the 15 most common fishes (>1% of total catch) were most abundant in rip-rap, two were most abundant in log jams, and none in mud banks. Rip-rap had the highest relative abundance of fluvial specialist and macrohabitat generalists, whereas mean CPUE of fluvial dependents was highest in log jams. Although a discriminant function analysis indicated that nine size classes (eight species) discriminated among three habitat types, the high misclassification rate (38%) suggested a high degree of fish assemblage overlap among the habitats. Although previous work has suggested that engineered structures (rip-rap) and urbanization are linked to reduced biotic diversity or reduced growth of fish species, our results suggest that at a local scale rip-rap may not have the same negative impacts on fish assemblages.