Benthic macroinvertebrates have been commonly used as indicator species for assessment of aquatic ecology. Streambed sediment, or substrate, plays an important role in habitat conditions for macroinvertebrate communities. Field investigations were done to study the benthic diversity and macroinvertebrate compositions in various stream substrata. Sampling sites with different bed sediment, latitude, and climate were selected along the Yangtze River, the Yellow River, the East River, and the Juma River, in China. The results show that benthic community structures found in different substrata clearly differ, while those found in substrata of similar composition and flow conditions but in different macroclimates are similar. The study, thus, demonstrates that the benthic macroinvertebrate community is mainly affected by substrate composition and flow conditions, but is generally unaffected by latitudinal position and macroclimate. Taxa richness of the macroinvertebrate community was found to be the highest on hydrophyte-covered cobbles, high on moss-covered bedrock, and low on clay beds and cobble beds devoid of plant biomass. Sandy beds are compact and unstable, thus, no benthic macroinvertebrates were found colonizing such substrata. Aquatic insects account for most of the macroinvertebrates collected in these rivers. Different insects dominate in different types of substrata: mainly EPT species (Ephemeroptera, Plecoptera, Trichoptera) in cobble, gravel, and moss-covered bedrock; and Chironomidae larvae in clay beds. The relation between the number of species in the samples and the size of the sampling area fits a power function of the species area. One square meter (1m2) is suggested as the minimum sampling area. A substrate suitability index is proposed by integrating the suitability of sediment, periphyton, and benthic organic materials for macroinvertebrates. The biodiversity of macroinvertebrates increases linearly with the substrate suitability index. Benthic taxa richness increases linearly with the suitability index.