The minimum number of thermoanalytical experiments that should be considered with the aim of performing a complete and reliable kinetic analysis when using semiempirical models is the problem of concern. It is shown by means of a series of numerical experiments that three differential scanning calorimetric (DSC) dynamic runs performed at different heating rates provide a complete and reliable kinetic analysis while the topological structure of two DSC curves is enough to determine the semiempirical model (among reaction order and SestackBergrenn) that better describe the experimental data. The procedure is analyzed by means of the simulation of a complex kinetic scheme and the results verified considering the thermal decomposition of cumene hydroperoxide. The proposed approach can be easily generalized to different kinetic models and is provided with a possible criterion devoted to identify autocatalytic processes by means of dynamical DSC experiments. (C) 2011 American Institute of Chemical Engineers AIChE J, 58: 18691879, 2012