Ozone tropospheric degradation of organic compound is very important in environmental chemistry. The lifetime of organic chemicals in the atmosphere can be calculated from the knowledge of the rate constant of their reaction with free radicals such as OH and NO3 or O3. In the present work, the rate constant for the tropospheric degradation of 137 organic compounds by reaction with ozone, the least widely and successfully modeled degradation process, are predicted by quantitative structure activity relationships modeling based on a variety of theoretical descriptors, which screened and selected by genetic algorithm variable subset selection procedure. These descriptors which can be used as inputs for generated artificial neural networks are; HOMO–LUMO gap, number of double bonds, number of single bonds, maximum net charge on C atom, minimum (>0.1) bond order of C atom and Minimum e–e repulsion of H atom. After generation, optimization and training of artificial neural network, network was used for the prediction of log KO3 for the validation set. The root mean square error for the neural network calculated log KO3 for training, prediction and validation set are 0.357, 0.460 and 0.481, respectively, which are smaller than those obtained by multiple linear regressions model (1.217, 0.870 and 0.968, respectively). Results obtained reveal the reliability and good predictivity of neural network model for the prediction of ozone tropospheric degradations rate constant of organic compounds
