Unionid freshwater mussels are one of the most imperilled groups in North America. They play an important role in freshwater ecosystems, both as a food source and as filter feeders. Their priority conservation status has generated interest in unionid research. Here, data from the US Geological Survey was used to produce predictive models of mussel habitat affinities at a resolution of 100 m2 across an area of thousands of square kilometres. This approach correctly identifies areas that are more suitable for threatened mussel species beds as compared with less suitable areas (>97% of the time) Stream segments were identified that are forecast to have high suitability for threatened mussels. Potamilus amphichaenus differed from other threatened mussel species by being associated with a wider range of volumetric flow rates and by not being restricted by the clay content of the soils. Of the species examined, it was the most large‐river oriented in habitat use and distribution. These methods can help conservation planners and land‐use managers make rational decisions about where to focus their efforts in lotic habitats without the need for intensive environmental measurements while still providing high‐resolution information.