We have constructed a composite image of the fault systems of the M 6.7 San Fernando (1971) and Northridge (1994), California, earthquakes, using industry reflection and oil test well data in the upper few kilometers of the crust, relocated aftershocks in the seismogenic crust, and LARSE II (Los Angeles Region Seismic Experiment, Phase II) reflection data in the middle and lower crust. In this image, the San Fernando fault system appears to consist of a decollement that extends 50 km northward at a dip of ∼25° from near the surface at the Northridge Hills fault, in the northern San Fernando Valley, to the San Andreas fault in the middle to lower crust. It follows a prominent aseismic reflective zone below and northward of the main-shock hypocenter. Interpreted upward splays off this decollement include the Mission Hills and San Gabriel faults and the two main rupture planes of the San Fernando earthquake, which appear to divide the hanging wall into shingle- or wedge-like blocks. In contrast, the fault system for the Northridge earthquake appears simple, at least east of the LARSE II transect, consisting of a fault that extends 20 km southward at a dip of ∼33° from ∼7 km depth beneath the Santa Susana Mountains, where it abuts the interpreted San Fernando decollement, to ∼20 km depth beneath the Santa Monica Mountains. It follows a weak aseismic reflective zone below and southward of the main-shock hypocenter. The middle crustal reflective zone along the interpreted San Fernando decollement appears similar to a reflective zone imaged beneath the San Gabriel Mountains along the LARSE I transect, to the east, in that it appears to connect major reverse or thrust faults in the Los Angeles region to the San Andreas fault. However, it differs in having a moderate versus a gentle dip and in containing no mid-crustal bright reflections.