Fluid injection in the subsurface has significantly increased over the last decades. Fluid injection causes pressure buildup, which reduces effective stresses and thus, shear failure conditions may eventually occur, inducing microseismic and seismic events. Anticipating felt induced earthquakes that may be triggered in undetected faults is crucial for the success of fluid injection projects. We propose a methodology to detect and locate such low-permeability faults to reduce the risk of inducing felt seismic events. The methodology consists in using diagnostic plots to identify the divergence time between the logarithmic derivative of overpressure evolution measured in the field and the one that would correspond to an aquifer including the previously identified heterogeneities. We apply the proposed methodology to water and CO2 injection through a horizontal well in a confined aquifer that has faults parallel to the well. We numerically obtain type curves that allow locating low-permeability faults once the divergence time is determined from the logarithmic derivative of overpressure for both water and CO2 injection. Furthermore, we illustrate how to apply the methodology to detect multiple low-permeability faults. To perform a proper pressure management based on fault stability analysis to minimize the risk of inducing felt seismic events, the proposed methodology should be complemented with other multidisciplinary tools. This methodology can be extended to other geological settings and be used in several fluid injection applications.