A method for determining the fault location of a ground fault relating to a line of a three-phase electrical energy supply network with an ungrounded neutral point, in which current and voltage values are measured at one line end of the line. Following the occurrence of a ground fault on the line, the location of the fault is defined using the measured current and voltage values. The measured current and voltage values are converted into α-components and 0-components of a Clarke transformation, a location-dependent characteristic function of α-component and 0-component fault voltage values is defined in each case using the converted components and a propagation model for travelling waves on the line. The fault location is determined to be the location at which the two characteristic functions exhibit the closest match.