Simulation data provided in this paper revealed that when a receptor full agonist and a competitive or noncompetitive antagonist for the same type of receptor are covalently linked together in some proper ways, the resulting bifunctional compound would have a reduced efficacy, and their dose–response curve patterns would look exactly like the curve patterns for many known partial agonists. It is also shown that all known pharmacologic characteristics for receptor partial agonists matched precisely the projected properties of the agonist–antagonist bifunctional compounds. The novel mechanistic explanation developed in this paper not only reveals the structural requirements for receptor partial agonists, but it also negates the long-held mechanistic explanation that the reduced efficacy of receptor partial agonists is the result of partial activation of each individual receptor molecule being occupied by the partial agonist because of its “low intrinsic activity”.