Abstract: The opening-mode (mode-I) dynamic fracture experiment for a sandstone was performed by using single cleavage drilled compression (SCDC) specimens impacted by split Hopkinson pressure bar (SHPB). The dynamic fracture toughness of rock was determined by two methods respectively. The first one was the experimental-numerical method: the dynamic load history was obtained from recordings of strain gauges glued on the bars of SHPB, the load was inputted to the finite element analysis program to calculate the dynamic stress intensity factor at the crack tip of specimens, and the dynamic fracture toughness is the stress intensity factor at the time of dynamic fracture initiation; the second one was the quasi-static method: the maximum load was inputted to the formula of a static stress intensity factor to determine the dynamic fracture toughness. Large differences of results were found between the experimental-numerical method and the quasi-static method: for specimens with dimensionless crack length a/R=0.64 (group A), the results of the quasi-static method were smaller than the results of the experimental-numerical method by 35%~62%; for specimens with dimensionless crack length a/R=1.61 (group B), the results of the quasi-static method were smaller than the results of the experimental-numerical method by 72%~83%. In principle, the experimental-numerical method can be used to measure rock dynamic fracture toughness more reasonably than the quasi-static method.