A theoretical analysis has been developed to predict fluorescence nonradiative energy transfer (NBET) behavior in homogeneous and phase-separated polymer blends. Conditions where intermolecular correlations need to be included are examined by first investigating the effect of including intermolecular correlations in predictions of NRET behavior in donor and trap (acceptor) end-labeled polymer melts. Donor fluorescence decays and energy transfer efficiencies are predicted for several different polymer systems using donor-trap intermolecular correlations in the theoretical analysis. These results are compared quantitatively to the same predictions recalculated without correlations and demonstrate the need to consider the effects of correlations when analyzing NRET measurements used for quantitative study of phase behavior. For the nonradiative energy transfer systems investigated here, correlation effects can often result in substantial differences, up to 60% as compared to the uncorrelated case, in predictions of relative energy transfer efficiency for bulk polymer. In the case of the blends, the effect of including intermolecular correlations is strongly a function of composition. A two-phase model is proposed to establish a quantitative method for relating energy transfer efficiency to phase-separated blend composition, and it is demonstrated that significant errors in interpretation of experimental NRET data may result if correlation effects are not included. © 1994 John Wiley & Sons, Inc.