Diverse contaminants of emerging concern (CECs) can be found in nowadays aquatic environment, possessing high potential to cause adverse ecological and human health effects. Due to their recalcitrance, conventional water treatment methods are shown to be inadequately effective. Thus, their upgrade by advanced oxidation processes, involving the generation of highly reactive species (HO and SO4-), is highly demanded. In order to assess the susceptibility of CECs by HO and SO4-, as well as to determine the corresponding reaction rate constants kHO and kSO4-, the complex experimental studies has to be maintained. The alternative is the application of modeling approaches which correlate structural characteristics with activities/properties of interest, i.e. quantitative structure activity/property relationship (QSAR/QSPR). In this study kHO and kSO4- of fifteen selected CECs were determined by competitive kinetics, and afterward used to elucidate key structural features promoting their degradation. In that purpose, QSPR models were constructed using multiple linear regression (MLR) combined with genetic algorithm (GA) approach. The models were submitted to the internal and external validation (using additional set of 17 CECs). Selected 3-variable models predicting kHO and kSO4- were characterized with high accuracy and predictivity (R2 = 0.876 and Q2 = 0.847 and R2 = 0.832 and Q2 = 0.778, respectively). Although selected models at the first sight include descriptors derived through complicated calculation procedures, their weighting schemes indicate on their relevance and transparency toward established reaction theories and differences regarding radical type.