Ornithomimus (/ˌɔːrnɪθəˈmaɪməs, -θoʊ-/; "bird mimic") is a genus of ornithomimid theropod dinosaurs from the Campanian and Maastrichtian ages of the Late Cretaceous in western North America. Ornithomimus was a swift, bipedal dinosaur which was covered in feathers and equipped with a small toothless beak that may indicate an omnivorous diet. It is usually classified into two species: the type species, Ornithomimus velox, and a referred species, Ornithomimus edmontonicus.

O. velox was named in 1890 by Othniel Charles Marsh on the basis of a foot and partial hand from the Denver Formation of Colorado. Other seventeen species have been named since then, though almost all of them have been subsequently assigned to new genera or shown not to be directly related to O. velox. The best material of species still considered part of the genus has been found in Alberta, representing the species O. edmontonicus, known from several skeletons from the Horseshoe Canyon Formation. Additional species and specimens from other formations are sometimes classified in the genus, such as Ornithomimus samueli (alternatively classified in the genera Dromiceiomimus or Struthiomimus) from the earlier Dinosaur Park Formation.

The history of Ornithomimus classification and the classification of ornithomimids in general has been very complicated. The type species, Ornithomimus velox, was first named by O.C. Marsh in 1890 and is based on syntypes YPM 542 and YPM 548 (a partial hindlimb and forelimb, respectively), found by George Lyman Cannon in the Denver Formation of Colorado on June 30, 1889. The generic name means "bird mimic", derived from Greek words ὄρνις (ornis), "bird", and μῖμος (mimos), "mimic", in reference to the bird-like foot. The specific name means "swift" in Latin. Simultaneously, Marsh named two other species: Ornithomimus tenuis (based on specimen USNM 5814) and Ornithomimus grandis. Both consist of fragmentary fossils found by John Bell Hatcher in Montana, which is today understood as tyrannosauroid material. At first, Marsh assumed Ornithomimus was an ornithopod, but this changed when Hatcher found specimen USNM 4736, a partial ornithomimid skeleton, in Wyoming. Marsh named it Ornithomimus sedens in 1892. On that occasion, Ornithomimus minutus was also created based on specimen YPM 1049 (a metatarsus), but it has since been recognized as belonging to an alvarezsaurid.

A sixth species, Ornithomimus altus, was named in 1902 by Lawrence Lambe and was based on specimen CMN 930 (hindlimbs found in 1901 in Alberta), but this was renamed to a separate genus in 1916: Struthiomimus, by Henry Fairfield Osborn. In 1920, Charles Whitney Gilmore named Ornithomimus affinis for Dryosaurus grandis (Lull 1911), based on indeterminate material. In 1930, Loris Russell renamed Struthiomimus brevetertius (Parks 1926) and Struthiomimus samueli (Parks 1928) into Ornithomimus brevitertius, Ornithomimus mirifica and Ornithomimus samueli, respectively. The very same year, Oliver Perry Hay renamed Aublysodon mirandus (Leidy 1868) into Ornithomimus mirandus, which is today seen as a nomen dubium. In 1933, William Arthur Parks created the species Ornithomimus elegans, which is today seen as either Chirostenotes or Elmisaurus. That same year, Gilmore named Ornithomimus asiaticus for material found in Inner Mongolia.

Also in 1933, Charles Mortram Sternberg named the species Ornithomimus edmontonicus for a nearly complete skeleton from the Horseshoe Canyon Formation (specimen CMN 8632).

At first, it had been common practice to name each newly discovered ornithomimid as a species of Ornithomimus. In the sixties, this tendency was still very strong, as is shown by the fact that Oskar Kuhn renamed Megalosaurus lonzeensis (Dollo 1903) from Belgium into Ornithomimus lonzeensis (which is understood today to be an abelisauroid claw) and Dale Russell in 1967 renamed Struthiomimus currellii (Parks 1933) and Struthiomimus ingens (Parks 1933) into Ornithomimus currellii and Ornithomimus ingens, respectively. At the same time, it was usual that workers referred to the entire ornithomimid material as simply "Struthiomimus". To solve this confusion by scientifically testing the separation between Ornithomimus and Struthiomimus, Dale Russell in 1972 published a morphometric study. It showed that statistical differences in some proportions could be used to distinguish the two and he concluded that Struthiomimus and Ornithomimus were valid genera. In the latter, Russell recognised two species: the type species Ornithomimus velox and Ornithomimus edmontonicus (even though he had trouble reliably distinguishing it from O. velox). He considered Struthiomimus currellii to be a younger synonym of Ornithomimus edmontonicus. However, Russell also interpreted the data as indicating that many specimens could not be referred to either Ornithomimus or Struthiomimus. Therefore, he created two new genera. The first one was Archaeornithomimus. Ornithomimus asiaticus and Ornithomimus affinis were reassigned to this new genus, becoming Archaeornithomimus asiaticus and Archaeornithomimus affinis. The second one was Dromiceiomimus, meaning "Emu mimic". This comes from the old generic name for the emu: Dromiceius. Russell assigned several former Ornithomimus species named during the 20th century, including O. brevitertius and O. ingens, to this new genus as Dromiceiomimus brevitertius. He also renamed Ornithomimus samueli into a second Dromiceiomimus species: Dromiceiomimus samueli.

Two tibiae from the Navesink Formation of New Jersey were named Coelosaurus antiquus ("antique hollow lizard") by Joseph Leidy in 1865. The tibiae were first attributed to Ornithomimus in 1979 by Donald Baird and John R. Horner as Ornithomimus antiquus. Normally, this would have made Ornithomimus a junior synonym of Coelosaurus, but Baird and Horner discovered that the name "Coelosaurus" was preoccupied by a dubious taxon, which was based on a single vertebra. It was originally named Coelosaurus by an anonymous author now known to be Richard Owen in 1854. Baird referred several other specimens from New Jersey and Maryland to O. antiquus. Beginning in 1997, Robert M. Sullivan regarded O. velox and O. edmontonicus as junior synonyms of O. antiquus. Like Russell, he considered the former two species indistinguishable from each other and noted that they both shared distinctive features with O. antiquus. However, David Weishampel (2004) considered "C." antiquus to be indeterminate among ornithomimosaurs, resulting in it being a nomen dubium. An SVP 2012 abstract agreed with Weishampel by noting that Coelosaurus differs from Gallimimus and Ornithomimus in the features of the tibiae.

In 1988, Gregory S. Paul classified the various species of Archaeornithomimus, Struthiomimus, Dromiceiomimus, and Gallimimus to the genus Ornithomimus. This has found no acceptance among other workers and the name is not presently used by Paul himself.