Secodontosaurus (meaning "cutting-tooth lizard") is an extinct genus of "pelycosaur" synapsids that lived from between about 285 to 272 million years ago during the Early Permian. Like the well known Dimetrodon, Secodontosaurus is a carnivorous member of the Eupelycosauria family Sphenacodontidae and has a similar tall dorsal sail. However, its skull is long, low, and narrow, with slender jaws that have teeth that are very similar in size and shape—unlike the shorter, deep skull of Dimetrodon ("two-measure tooth"), which has large, prominent canine-like teeth in front and smaller slicing teeth further back in its jaws. Its unusual long, narrow jaws suggest that Secodontosaurus may have been specialized for catching fish or for hunting prey that lived or hid in burrows or crevices. Although no complete skeletons are currently known, Secodontosaurus likely ranged from about 2 to 2.7 metres (7–9 ft) in length, weighing up to 110 kilograms (250 lb).

Fossils of Secodontosaurus have been found in Texas in North America in the Wichita and the Clear Fork groups of Early Permian formations. In recent years, teams from the Houston Museum of Natural Science have recovered remains in the Clear Fork Red Beds of North Texas that appear to be new specimens of Secodontosaurus. These discoveries are mentioned in online blogs but so far have not been formally described.

The name Secodontosaurus comes from Latin seco ('to cut') + Greek ὀδούς, ὀδόντος (odoús, -odont, 'tooth') + Greek σαῦρος (-saurus, 'lizard') and is based on the anatomical term "secodont" for teeth with cutting edges designed to tear or slice flesh. Paleontologist Robert Bakker has dubbed Secodontosaurus the "fox-faced finback" after its long jaws.

A number of partial fossil remains of Secodontosaurus have been identified from its characteristic long skull and jaws. The postcranial skeletal material from different individuals includes parts of the backbone with clear evidence of a tall sail very similar to that of Dimetrodon. The limbs and tail are incomplete but probably resembled those of Dimetrodon as well. Like Dimetrodon, Secodontosaurus would have had a short neck, robust body, short limbs, and a long tail. In light of such similarities, some skeletal remains with missing or fragmentary skulls that were previously identified as Dimetrodon may in fact belong to Secodontosaurus. A key noncranial difference can be found in the axis neck vertebra, which has a tall and broad neural spine in Dimetrodon but has a lower neural spine in Secodontosaurus.

Robert R. Reisz and others described the skull in detail in 1992, based mostly on a nearly complete skull specimen (MCZ 1124) about 27 cm (11 in) long, preserved with a left mandible. In addition to the long, low skull and nearly uniform size of the teeth that contrast with Dimetrodon, the anterior teeth of the upper jaw are slanted back and those of the lower jaw are directed forward for grasping prey. Its rather crocodile-like skull suggests that Secodontosaurus could have been semi-aquatic and may have fed on fish and small swimming amphibians. However, Reisz and his coauthors noted that a tall sail would seem to be a hindrance in pursuit of quick-moving creatures underwater. Instead, the long, narrow snout with forward slanting teeth at the mandible tip might have allowed Secodontosaurus to probe after small animals hiding in burrows and other tight spaces.

The American paleontologist Edward Drinker Cope published the first description of Secodontosaurus material in 1880 as a supposed species of his genus Theropleura ("mammal rib"). Theropleura Cope, 1878 is a junior synonym of Ophiacodon Marsh, 1878 ("snake point tooth"). The specimen, AMNH 4007 collected by Jacob Boll, was actually a composite that included bones of "amphibians" and a sphenacodontid. Cope described the teeth as having "apices are not very acute. The superficial coating is striate with fifteen or sixteen rather obtuse ridges"—features expressed in the species name Theropleura obtusidens (Latin for "obtuse tooth" or "blunt-toothed").

E. C. Case described another specimen of Secodontosaurus, AMNH 4091, as Dimetrodon longiramus ["long (lower jaw) ramus"] in 1907.

In 1916 S. W. Williston illustrated parts of the upper and lower jaws of a new, then unnamed genus, specimen FMNH (WM) 573, that he took for a member of the ophiacodontids, noting, however, its "broader, flattened, and cutting teeth" compared to Ophiacodon, which has mainly pointed conical teeth. In his 1925 Osteology of the Reptiles, Williston listed the name Secodontosaurus under the Ophiacodontidae, presumably as the new genus from 1916, but did not provide a description or an explanation. Williston likely chose the name Secodontosaurus "cutting-tooth lizard" for flat, sharp-edged ["secodont"] teeth that contrasted with the conical teeth typically found in ophiacodontids.