A weed survey was conducted on 134 Palmer amaranth (Amaranthus palmeri S. Watson) populations from Mississippi and Arkansas in 2017 to investigate the spread of resistance to protoporphyrinogen oxidase (PPO) inhibitors using fomesafen as a proxy. Fomesafen resistance was found in 42% of the A. palmeri populations. To investigate the resistance basis of different PPO inhibitors, we further characterized 10 representative populations by in planta bioassay in a controlled environment and molecular characterizations (DNA sequencing and TaqMan^® gene expression assay). A total of 160 plants were sprayed with a labeled field rate (1X) of fomesafen or salfufenacil and screened for the presence of three known resistance-endowing mutations in the mitochondrial PPX2 gene (ΔGly-210, Arg-128-Gly, Gly-399-Ala). To compare the potencies of fomesafen and saflufenacil, dose–response studies were conducted on two highly resistant and one sensitive populations. The interaction of the two herbicides with the target protein harboring known PPX2 mutations was also analyzed. Our results showed that: (1) 90% of the fomesafen- or saflufenacil-resistant plants have at least one of the three known PPX2 mutations, with ΔGly-210 being the most prevalent; (2) saflufenacil is more potent than fomesafen, with five to nine times lower resistance/susceptible (R/S) ratios; (3) fomesafen selects for more diverse mutations, and computational inhibitor/target modeling of fomesafen suggest a weaker binding affinity in addition to a smaller interaction volume and volume overlap with the substrate protoporphyrinogen IX than saflufenacil. As a result, saflufenacil shows reduced sensitivity to PPX2 target-site mutations. Results from current study can help pave the way for designing weed management strategies to delay resistance development and maintain the efficacy of PPO inhibitors.