Diffuse-field theory is used by practitioners to predict sound fields in rooms of every type. Often forgotten is the fact that the theory is based on assumptions which may limit its applicability. If the theoretical assumptions do not hold in the case of a particular room for which predictions are to be done, the predictions may not be accurate. The objective of this paper is to review what is known about the applicability of diffuse-field theory. This is mainly based on extensive work by Kuttruff and by the author, comparing predictions by diffuse-field theory and by ray-tracing models. It considers two versions of diffuse-field theory—the Eyring and Sabine versions—and the prediction of both sound decay/reverberation time, and of steady-state sound pressure level. It discusses applicability with respect to the following room-acoustical parameters: room shape; surface absorption (spatial distribution and magnitude); surface reflection; fitting density. How the results apply to real rooms is discussed and illustrated using experimental results.