California has experienced unprecedented wildfire activity in recent years, as half of the twenty largest fires in the state's recorded history occurred during 2020-2021. Large and severe wildfires have substantial potential to degrade or effectively eliminate, for the near future, habitat of forest-dependent species such as martens or fishers. For instance, >33% of predicted habitat for the endangered southern Sierra fisher population burned in the past 2 years and >55% since 2012. We describe conditions in the Sierra Nevada with >10 years of animal demographic data in 2 locations. In the northern Sierra, we estimated the influence of fire on predicted habitat and connectivity for marten. We created predicted habitat models from extensive location data and spatially optimized variables. We evaluated models using cross validation and independent data, then created a resistance surface (negative exponential predicted habitat quality). The Dixie fire burned 45% of marten predicted habitat cores and 38% of corridors, 24% of which were at moderate or high severity. In both study areas we ran new simulations following fire, assuming high resistance in moderate/high severity burned patches to identify areas for optimized restoration planning. We conclude with creative ideas and possibilities for maintaining forest dependent species in landscapes with prior and ongoing megafires.