The invention relates to a method of locating in a dispersive medium (5) a source (7) of pulses (s1, s2) by at least one pair of sensors (1, 2), the method comprising the following steps: at the level of each sensor (1, 2), detecting a pulse (s1, s2) originating from the source (7); for each of said pulses, constructing a phase diagram on the basis of N vectors of which the co-ordinates correspond to the amplitude of the pulse (s1, s2) at successive sampling instants ti; and for each pair of sensors, calculating the ratio between the distances L1 and L2 of each sensor from the source through the formula (I) where (II) and (III) are the norms of the vectors (IV) and (V) of the phase diagrams corresponding to the pulses detected by the sensors (1, 2).