Quantitative structure-property relationships(QSPRs) have been developed to predict the thermal stability for a set of 22 nitroaromatic compounds by means of the theoretical descriptors derived from electrostatic potentials on molecular surface. Several techniques, including partial least squares regression(PLS), least-squares support vector machine(LSSVM) and Gaussian process(GP) have been utilized to establish the relationships between the structural descriptor and the decomposition enthalpy. The nonlinear LSSVM and GP models have proven to own a better predictive ability than the linear PLS method. Moreover, owing to its ability to handle both linear- and nonlinear-hybrid relationship, GP gives a stronger fitting ability and a better predictive power than LSSVM, and therefore could be well applied to developing QSPR models for the thermal stability of nitroaromatic explosives. Quantitative structure-property relationships (QSPRs) have been developed to predict the thermal stability for a set of 22 nitroaromatic compounds by means of the theoretical descriptors derived from electrostatic potentials on molecular surfaces. Several techniques, including partial least squares regression (PLS), least The -squares support vector machine (LSSVM) and Gaussian process (GP) have been utilized to establish the relationships between the structural descriptor and the decomposition enthalpy. The nonlinear LSSVM and GP models have proven to own a better predictive ability than the linear PLS method. Moreover, owing to its ability to handle both linear- and nonlinear-hybrid relationship, GP gives a stronger fitting ability and a better predictive power than LSSVM, and therefore could well apply to developing QSPR models for the thermal stability of nitroaromatic explosives.