Leptictidium is a special animal because of the way its anatomy combines quite primitive elements with elements which prove a high degree of specialization. It had small fore legs and large hind legs, especially at the distal side (that further from the body). The lateral phalanges of its forelegs (fingers I and V) were very short and weak, finger III was longer and fingers II and IV were roughly equal in size, and slightly shorter than finger III. The tips of the phalanges were elongated and tapered.
The ankles and the sacroiliac joint were quite loosely fixed, while the pelvis had a flexible joint with only one coccygeal vertebra. The anteorbital muscle fenestrae in their crania suggest they probably had a long and mobile snout, similar to that of elephant shrews.
Leptictidium had wide diastemata in the antemolar row, its upper molar teeth were more transverse than those of the North American leptictids and its fourth premolars were molariform. Its C1 canines were incisiviform. Its dentition was quite small in comparison to the size of the mandible and the animal as a whole.
It varied between 60 and 90 cm (24–36 in) in length (more than half of which belonged to the bald tail), and 20 cm (8 in) in height. It weighed a couple of kilograms. These sizes could vary from one specimen to another.
One of the mysteries about Leptictidium is whether it moved by running or by jumping. Because there are very few completely bipedal mammals, it is difficult to find an appropriate living model to compare it with. If the kangaroo is used, it is probable that Leptictidium hopped along with its body tilted forward, using its tail as a counterweight. On the other hand, elephant shrews combine both types of locomotion; they usually move on four legs, but they can run on two legs to flee from a predator. Studies of the bone structure of Leptictidium have yielded contradicting information: its leg articulations appear too weak to have supported the shock of repeated jumps, but its long feet were obviously adapted for jumping rather than running.
Kenneth D. Rose compared the species L. nasutum with the leptictid Leptictis dakotensis. L. dakotensis had a series of traits which show it was a running animal which sometimes moved by jumping. Despite the marked similarities between Leptictis and Leptictidium, there are certain differences in their skeletons which prevent the example of Leptictis from being used to determine with certainty the way Leptictidium moved: the most important being that, unlike Leptictis, the tibia and the fibula of Leptictidium were not fused together.
Perfectly preserved fossils of three different species of Leptictidium have been found in the Messel pit in Germany. The marks on their fur have been preserved, as well as their stomach contents, which reveal Leptictidium were omnivores which fed on insects, lizards and small mammals. The holotype of L. tobieni also had pieces of leaves and notable amounts of sand in its abdomen, but it cannot be determined with certainty if the animal swallowed it.
Since Leptictidium young were very vulnerable to attacks of predators, it is probable that they were born quite developed and that they became independent when they were still quite young. From the behavior of today's elephant shrew, the possibility can be suggested that Leptictidium young went out to hunt with their mother, and that they learned to distinguish prey by licking their mother's mouth when she had just caught an animal, in order to know the flavor of food.