The July 2019 Ridgecrest earthquakes in southeastern California were characterized as surprising by some, because only of the rupture occurred on previously mapped faults. Employing more detailed inspection of pre‐event high‐resolution topography and imagery in combination with field observations, we document evidence of active faulting in the landscape along the entire fault system. Scarps, deflected drainages, and lineaments and contrasts in topography, vegetation, and ground color demonstrate previous slip on a dense network of orthogonal faults, consistent with patterns of ground surface rupture observed in 2019. Not all of these newly mapped fault strands ruptured in 2019. Outcrop‐scale field observations additionally reveal tufa lineaments and sheared Quaternary deposits. Neotectonic features are commonly short (), discontinuous, and display en echelon patterns along both the M 6.4 and M 7.1 ruptures. These features are generally more prominent and better preserved outside the late Pleistocene lake basins. Fault expression may also be related to deformation style: scarps and topographic lineaments are more prevalent in areas where substantial vertical motion occurred in 2019. Where strike‐slip displacement dominated in 2019, the faults are mainly expressed by less prominent tonal and vegetation features. Both the northeast‐ and northwest‐trending active‐fault systems are subparallel to regional bedrock fabrics that were established as early as , and may be reactivating these older structures. Overall, we estimate that 50%–70% (i.e., an additional 15%–35%) of the 2019 surface ruptures could have been recognized as active faults with detailed inspection of pre‐earthquake data. Similar detailed mapping of potential neotectonic features could help improve seismic hazard analyses in other regions of eastern California and elsewhere that likely have distributed faulting or incompletely mapped faults. In areas where faults cannot be resolved as single throughgoing structures, we recommend a zone of potential faulting should be used as a hazard model input.