The focus of the proposed effort is on the prediction of the fatigue life of composite panels subjected to thermal effects, inducing buckling, and a strong acoustic excitation which may force the panel to “jump” from one buckled configuration to another. The fatigue life will be estimated through the modeling of the probability density function of stress ranges from which the fatigue life is readily obtained. The proposed model is an extension of Dirlik’s formula that also accounts for the nonlinearity of the panel dynamics as well as the nonlinearity of the corresponding displacements-stresses relation. The proposed model which is theoretically justified will be validated/fine-tuned on an extensive database of stress time histories corresponding to 4 different structural models exhibiting a variety of spectral features, i.e. one or several peaks, one or several dominant peaks, close frequencies, etc., subjected to 4 different thermal loading scenarios, at various sound pressure levels, and with several different angle of incidence. The determination of the parameters of the model from either experimental time histories of the stresses or a finite element model of the panel will also be addressed. Although the proposed work addresses specifically the fatigue life of panels subjected to thermoacoustic effects, it is expected that the modified Dirlik formula would be applicable to a broad range of nonlinear vibration problems.