The management and recovery of populations of bull trout Salvelinus confluentus requires a comprehensive understanding of habitat use across different systems, life stages, and life history forms. To address these needs, we collected microhabitat use and availability data in three fluvial populations of bull trout in eastern Oregon. We evaluated diel differences in microhabitat use, the consistency of microhabitat use across systems and size-classes based on preference, and our ability to predict bull trout microhabitat use. Diel comparisons suggested bull trout continue to use deeper microhabitats with cover but shift into significantly slower habitats during nighttime periods; however, we observed no discrete differences in substrate use patterns across diel periods. Across life stages, we found that both juvenile and adult bull trout used slow-velocity microhabitats with cover, but the use of specific types varied. Both logistic regression and habitat preference analyses suggested that adult bull trout used deeper habitats than juveniles. Habitat preference analyses suggested that bull trout habitat use was consistent across all three systems, as chi-square tests rejected the null hypotheses that microhabitats were used in proportion to those available (P < 0.0001). Validation analyses indicated that the logistic regression models (juvenile and adult) were effective at predicting bull trout absence across all tests (specificity values = 100%); however, our ability to accurately predict bull trout absence was limited (sensitivity values = 0% across all tests). Our results highlight the limitations of the models used to predict microhabitat use for fish species like bull trout, which occur at naturally low densities. However, our results also demonstrate that bull trout microhabitat use patterns are generally consistent across systems, a pattern that parallels observations at both similar and larger scales and across life history forms. Thus, our results, in combination with previous bull trout habitat studies, provide managers with benchmarks for restoration in highly degraded systems.