The optimization of the HPLC procedures and the investigation of the predictive models are generally seen as distinct tasks although they could be both approached by very similar chemometric methods. In this work we used chemometric approaches to both optimize the separation of a non-steroidal anti-inflammatory drugs (NSAIDs) set by isocratic reversed-phase HPLC and start-up the development of retention predictive models. The screening of responsive variables and the search for an optimal experimental domain provided also an insight into the structure-chromatographic response relationships for the considered compounds which were particularly helpful to derive new retention parameters.