The rate constant for the nighttime degradation of 114 heterogeneous organic compounds, through reaction with nitrate radicals in the troposphere, is predicted here by quantitative structure-activity relationships modelling. The multiple linear regression approach is based on a variety of theoretical molecular descriptors, selected by the genetic algorithms-variable subset selection procedure. The proposed model, calculated on a limited subset of compounds selected by a D-optimal experimental design and checked for reliability and robustness, has good predictivity, verified by internal (QLMO 2=89.6%) and "external" validation (QEXT 2=95.7%). The model applicability domain was always verified by the leverage approach in order to propose reliable predicted data. The average root-mean square error for the prediction of logkNO3 was 0.57, similar to (and even smaller than) the typical experimental error range.