Introduction Compared with conventional construction methods in terms of the building structure, 3D-printing concrete (3DPC) technology has the advantages of formwork-free, low labor costs, and less construction waste. The integration of digital design and automatic construction improvs the construction efficiency. 3DPC technology based on extrusion manufacturing is applied in various scenarios such as bridges, landscape sketch, and buildings. However, the stacking-up method also causes some issues like layered appearance, mechanical anisotropy, and high requirements of constructability. At present, the mechanical properties of 3DPC are investigated experimentally, however, the researches on the mechanical properties of 3DPC mainly focus on the static properties of materials, and a few studies involve the dynamic mechanical properties of 3DPC. For concrete structures, they bear static loads, and inevitably face dynamic load (i.e., earthquakes). As a typical dynamic load, the strain rate of concrete under earthquakes ranges from 10-4 s-1 to 10-2 s-1, causing a catastrophic damage to engineering structures. It is essential to understand the dynamic mechanical properties of 3DPC for engineering application. In this paper, the dynamic mechanical properties of 3DPC in three orthogonal directions were investigated by uniaxial compression tests, and the influences of strain rate and different directions on the compressive strength, elastic modulus, peak strain, and anisotropy of 3DPC were analyzed.Methods A 42.5 fast hardening early strength sulfoaluminate cement and silica fume were used as cementitious materials. A local river sand was used as a fine aggregate. A polycarboxylate based water reducing agents with a water reducing rate of > 30% and a tartaric acid retarder were used as additives. A water/cement ratio of 3DPC was 0.29. A concrete printing system was applied to print blocks with dimensions of 220 mm×240 mm×400 mm at an extrusion speed of 1 L/min and a printing speed of 30 mm/s. The printed filament had a width of 30 mm and a height of 10 mm. The moving direction of printing nozzle on the horizontal plane is defined as x direction, where y direction is perpendicular to x direction, and z direction is the gravity direction and perpendicular to the x-y plane. After 28-d standard curing, cylinder specimens were obtained via taking cores from three directions of printed block, and then they were cut into 100 mm×200 mm specimens for compressive strength tests. An electro-hydraulic servo universal testing machine was used to applied dynamic loads. The applied strain rates were 10-5, 10-4, 10-3.5 s-1, and 10-3 s-1.Results and discussion When the weak interface is parallel to the loading direction, 3DPC is prone to failure along the interface due to the development of microcracks at the weak interface. At strain rates of 10-5 s-1 and 10-4 s-1, vertical cracks develop along the inter-layer and inter-strip interfaces in the y and z directions. When the strain rate increases to 10-3.5 s-1 and 10-3 s-1, a diagonal crack appear in the y direction, while cracks in the z direction still appear in the inter-strip interfaces. As the strain rate increases, the deformation rate cannot meet the demand for dissipated energy, leading to an increase of damage in the matrix. Therefore, cracks appear in all three directions in the matrix at higher strain rates.The dynamic peak stress in each direction has a certain degree of strain rate dependence.The dynamic compressive strength in all directions significantly increases as the strain rate increases. The compressive strength in the y direction is more affected by strain rate, compared to that in the x and z directions. At the same strain rate, the compressive strength in the x direction (Fx) is larger than that of cast concrete (FC) in the z (Fz) and y (Fy) directions, indicating that 3DPC has significant anisotropic dynamic compressive strengths. During the elastic stage, pores and defects in concrete are compacted. Their deformation and quantity have a significant impact on the elastic modulus of concrete. When the specimens in the z and y directions are compressed, there are more interfaces perpendicular to the compression direction, and the static elastic modulus in these two directions are smaller. As a result, the growth factors of elastic modulus in these two directions are more sensitive to the strain rate. However, as the strain rate increases, the peak strains in the x, y and z directions stabilize at 0.002.At the strain rates from 10-5 s-1 to 10-3.5 s-1, the anisotropy index of 3DPC decreases rapidly with the increase of strain rate. Because more defects are compacted at high strain rates, the differences in material properties in various directions decrease gradually, that is, the anisotropy of printed concrete decreases gradually. Conclusions The failure of 3DPC originated from defects at the layers or strips interface parallel to the loading direction, which was the weak part of the printed concrete. Under dynamic loading, the compressive strength of 3DPC exhibited a strain rate-dependence in all the directions with Fx>FC>Fz>Fy, and the strain rate effect in the y-direction was dominant. The dynamic increasing factors of compressive strength and elastic modulus increased linearly with the logarithm of strain rate ratio. When the loading strain rate increased from 10-5 s-1 to 10-3 s-1, the dynamic strengths of 3DPC in the x, y, and z directions were increased by 33.47%, 49.88% and 32.90%, respectively, compared to the strength at a strain rate of 10?5 S1. The elastic modulus of printed concrete in the x, y, and z directions was increased by 15.63%, 40.53% and 40.68%, respectively. However, the dynamic strength and elastic modulus of cast concrete were only increased by 31.56% and 13.60%. When the applied strain rate increased from 10-5 s-1 to 10-3 s-1, the anisotropy coefficient decreased from 5.69 to 2.96. The anisotropy of 3D printed concrete decreased with the increase of strain rate, but the change range was small at a high strain rate.