Cyclin-dependent kinases (CDKs) regulate the cell division cycle, apoptosis, transcription, differentiation and many functions in the nervous system. The frequent deregulation of CDKs in cancers and in numerous other pathologies justifies the active search for chemical inhibitors capable of reversibly and selectively inhibiting this class of enzymes. Intensive screening of collections of natural and synthetic compounds has led to the identification of several families of ATP competitive CDK inhibitors. As the therapeutic potential of the most promising compounds is currently being evaluated in preclinical and clinical trials, their mechanism of action is still unclear. In particular, the real spectrum of their intracellular targets remains largely unknown. Determination of the selectivity of the compounds and identification of their intracellular targets constitute a prerequisite to understand their cellular effects and to improve their efficiency on a rational basis. The classical method for the determination of a compound's selectivity consists in testing the compound in a panel of purified kinases. However, the selectivity study is then restricted to the panel's enzymes. As a consequence, many, if not most other potential targets are not evaluated. As an alternative way to investigate the range of true targets of CDK inhibitors, we propose an affinity chromatography approach based on immobilized inhibitors. Briefly, the inhibitor is covalently bound to a resin and cellular extracts are batch loaded on this inhibitor matrix. After extensive washing, the bound proteins are resolved by SDS-PAGE and identified by microsequencing. In addition to confirming the interaction of CDK inhibitors with CDKs, this method has led to the identification of additional, sometimes unexpected, targets. We here illustrate the potential of this technique through a few examples.