During the early days of the Apollo project, it was generaly agreed that there were three methods for a lunar landing - Direct, Earth Orbit Rendezvous, and Lunar Orbit Rendezvous.
Direct Ascent was widely favored during the first few years, because it was the simplest in terms of flight profile, theory, engineering, and mechanics. However, there was no booster powerful enough to loft an entire spacecraft straight from the Earth to the Moon. Although seemingly an issue solved by research and development into a bigger rocket, every time engineers created a concept seemingly powerful enough, they would discover the weight of the vehicle/payload would be increased.
Earth Orbit Rendezvous was the second primary competitor against Direct Ascent. The vehicle would be sent into a low-earth orbit, whether through a single or multiple launches, where it would then be sent directly to the Lunar surface, and vice versa. Cost was a large factor in the dismissal of this mode, as NASA was unsure if the government or public would support multiple launches for a single mission.
Lunar Orbit Rendevouz was the final, most improbable choice. A space vehicle would be lofted directly to the moon, but instead of landing straight on the surface, it would first go into a parking orbit. The vehicle would then land on the surface, and return directly to the Earth. The least favored of the three, LOR would require multiple space feats that were at the time considered impossible, difficult at best, such as rendezvous, docking, restarting an engine et. al. This mode ultimately won when it was determined that the liftoff weight of the booster could be severely diminished if the vehicle achieved a low Earth parking orbit for systems checkout before injecting itself into a trans-lunar trajectory.