Various kinetic models, namely, the Avrami analysis modified by Jeziorny, the Ozawa model, and a method developed by Mo, were applied to describe the nonisothermal melt crystallization process of poly(ethylene terephthalate) (PET) and two PET/clay nanocomposites of different viscosities. The Avrami analysis modified by Jeziorny could gratifyingly describe the primary nonisothermal crystallization stage of PET and two PET/clay nanocomposites. The difference in the values of the Avrami exponent between PET and PET/clay nanocomposites suggested that the nonisothermal crystallization of PET/clay nanocomposites corresponds to a tridimensional growth with heterogeneous nucleation. The values of half-time showed that the crystallization rate of PET/clay nanocomposites is faster than that of PET at a given cooling rate. The Ozawa analysis failed to provide an adequate description of the nonisothermal crystallization of PET/clay nanocomposites. The method developed by Me was successful in describing the nonisothermal crystallization of pristine PET and PET/clay nanocomposites. The activation energy for nonisothermal crystallization of pristine PET and two PET/clay nanocomposites of different viscosities, based on the Augis-Bennett method, the Kissinger method, and the Takhor method, respectively, were evaluated, and it was concluded that the absolute value of activation energy for PET is lower than that of PET/clay nanocomposites, and this showed that introduction of clay into PET matrix weakens the dependence of the nonisothermal crystallization exotherms peak temperatures on the cooling rates used. (C) 2003 Wiley Periodicals, Inc.