The Human Genome Project primarily used the Whole Genome Shotgun (WGS) sequencing approach, which involved randomly fragmenting the genome, sequencing fragments, and assembling them computationally. This was more efficient than Sanger or Maxam-Gilbert methods for large-scale genome sequencing.