We experimentally verified that material-averaging techniques can be used for numerically modeling the response of the crosswell electromagnetic (crosswell EM) system in a fractured medium. We have designed a scaled model of the crosswell EM system, and verified that the laboratory data are in good agreement with the simulated response for different cases of fractures present in a host medium. Simulations of realistic scenarios in a hydrocarbon-filled reservoir indicate that the presence of fracture clusters can significantly affect the crosswell EM system response. The magnitude of the effect primarily depends on the number of fractures and their dip relative to the transmitter and receiver wells. In particular, small clusters of fractures filled with saline connate water produce large artifacts on the response, provided that the relative dip to the wells is large enough. The presence of a fracture cluster creates a characteristic signature in the full crosswell EM tomographical data set; therefore, crosswell EM data provides information related to the main fractures that are present between the wells. The inclusion of the fracture information as an additional constraint to the inversion of the crosswell EM data is expected to provide an improved interpretation of the reservoir fluids in the interwell space.