Brown trout and young Atlantic salmon in streams are selective in their habitat use, which is partially determined by hydro‐physical conditions. Habitat selection may be quantified in models and combined with hydraulic models to evaluate instream habitat suitability. Fish occupancy of habitat depends on the fish species and size. Brown trout prefer deep stream areas with moderate to low water velocities and rocky substrates, whereas young Atlantic salmon chose more fast flowing and often shallower areas. Habitat selection has been quantified in static selection models which should be based on measures of habitat usage and availability (preferences) and combined with data on hydro‐physical conditions to build predictive habitat hydraulic models. Such models assess habitat availability and capacity rather than discharge–biomass relationships. Limitations of static models in fish habitat studies are (1) the relevant hydro‐physical variables are not included, (2) the interaction terms are difficult to quantify and not incorporated, (3) the hydraulic models may not operate on a spatial scale that is relevant to fish, (4) the models include spatial but only to a limited extent temporal heterogeneity in habitat conditions and (5) biotic factors are not included. Streams may be extremely heterogeneous ecosystems, both spatially and temporally, which may influence habitat selection and modelling. In response to varying habitat availabilities (stream size and structure, water flows) habitat selection in brown trout and young Atlantic salmon is dynamic and relatively flexible. Furthermore, changes in temperature may result in seasonal and daily niche shifts. Therefore unless the dynamic aspects of habitat selection are incorporated into the habitat models, long‐term predictive power in habitat–hydraulic modelling is unlikely. However, habitat–hydraulic modelling is a useful tool in a ‘no net loss of habitat’ management strategy regardless of these shortcomings.