Classical interpretation techniques developed primarily for gravity and magnetic data have been adapted to electric self-potential (SP) data. From a recently developed wavelet-based technique for potential field interpretation, we propose a novel approach for the forward and inverse modeling of SP anomalies caused by subsurface fluid flow. We use analytic signals and wavelets associated with the electric and fluid flow potentials. Fluid flow singularities have typical signatures in the wavelet-domain SP data that can be used in the inversion scheme. For cases where electrokinetics is unambiguously the origin of SP anomalies, our interpretation method provides key parameters of the subsurface fluid flow (e.g. position, geometry, flow rate). To illustrate the theory, we consider synthetic 2D flows for which analytic solutions exist and actual SP data acquired around the Vulcarolo fissure on Mount Etna that show sources between 3 to 18 m depths. ¿ 2001 American Geophysical Union