It is the purpose of this study to obtain the fundamental information on the fatigue behavior of wood and wood based materials under impact blows applied repeatedly on their bending specimens. Four kinds of test materials were used, namely, Taiwanhinoki (Chamaecyparis taiwanensis M. et S.), Hemlock (Tsuga heterophylla Sarg.), sevenlayer plywood and threelayer particleboard. The impact bending tests were carried out by center loading method on the dropping type impact machine. The weight of dropping hammer was 8.5 kg and the test specimens were made on 20 mm thick by 20, 40 or 60 mm wide by 320 mm long. The test results obtained are summarized as follows. The fracture types of specimens in the impact fatigue tests differ with each material. The type in the case of solid wood (Taiwanhinoki and Hemlock) is horizontal shear slit near the neutral axis of bending specimen and tension fracture at the center of bending span on the tensile side. The type in plywood is rolling and horizontal shear fracture on the veneer layers which intersect at right angle to lengthwise, or tension fracture on the outer layer, and these fractures are depended on their widths of the specimens. The type in particleboard is brush tension fracture at the center of span. There exists a negative rectilinear correlation on a log scale between the impact bending strength of test specimen (σ_b) and the repetition number of impact blow to fracture (N) in all materials. The values of impact energy applied to a test specimen (E_b) when solid wood and plywood are fractured by the first impact blow show about ten and five times as much as that of particleboard respectively. The strain distribution along the tensile side on the bending specimen when a impact blow is applied to it shows the curve which has a maximum value at the center of span and has a decreasing slope with going away from the center, and the decreasing slope is steep in the case of solid wood, but it is gentle in the case of particleboard. And plywood shows a medium slope of them. The maximum strain on the center of span at n repetition number of impact blow (ε_M) is presented by the sum of the strain generated by the impact blow itself (impact strain) and the strain remained by the last blow (residual strain), and the both strains increase with increase of n. The final value of ε_M when the specimen is fractured is depended on not only tested material but also N. In the case of solid wood, the value is constant regardless of N in the smaller range of N but it increases gradually with N above the range, and the value is largest among the tested materials. The variation of ε_M with N in the case of plywood resembles to the case of solid wood and the value follows the case. On the contrary of those, ε_M in the case of particleboard decreases with increase of N, and the value of ε_M is smaller than that of the other two materials. Hysteresis energy accumulated in the specimen by every impact blow increases with increase of n, and the accumulated values of the energy until the final blow in the case of solid wood and plywood are about forty-five and fifteen times as much as that of particleboard respectively.