We present a new method for evaluating network performance in surface and borehole seismic monitoring. For a specified network geometry, station characteristics and a target monitoring volume, the method determines the lowest magnitude of events that the seismic network is able to detect (Mwdetect) and locate (Mwloc), and estimates the expected location and origin time errors for a specified magnitude. Many features of the seismic signal recorded at a single station are considered, including characteristics of the seismic source, the instrument response, the ambient noise level, and wave propagation in a layered, anelastic medium using asymptotic ray-theory. In application of the methodology, P and S signal-to-noise ratio thresholds are defined and, based on synthetic waveforms, a count is made of the number of stations that can detect and pick P or S phase arrivals in the presence of noise. Then, Mwdetect and Mwloc are determined by the smallest magnitude events, which have a specified number of phase detections and phase picks, respectively. Finally, we map expected location errors using the prediction analysis formalism of Peters and Crosson (1972), extended to include P-wave polarization uncertainties (of both the azimuth and the inclination angles) in addition to velocity model uncertainties and the P- and S-wave arrival time uncertainties for each station of the network. We applied this method to two different network typologies: a local earthquake monitoring network, ISNet (Irpinia Seismic Network) installed along the Campania-Lucania Apennine chain in Southern Italy, and a hypothetic borehole network for monitoring micro-fractures induced during the hydrocarbon extraction process in an oil field. The method can be used to improve existing networks and understand their capabilities, such as for the ISNet case study, or to optimally design the network geometry in specific target regions, as for the borehole network example.