he observation of molecules in the interstellar medium gives information on the physic of the formation of stars and planets. For instance, each step of the formation of a planetary system is characterized by a chemical composition, which directly reflects the physical conditions of the medium and its evolutionary stage. Dense clouds show large abundances of unsaturated species formed in cold gas phase whereas in protostars, gas warm up favors grain surface reactions and the formation of large saturated molecules. With the improvement of observational instruments, more and more trace molecules are detected in the ISM. The future far-infrared space telescope Herschel and (sub) millimeter interferometer ALMA are promises of opening new windows on the ISM wealth. The more molecules are discovered, the more complex chemical models have to be to reproduce this peculiar chemistry. For the future, there are two main challenges to take up. The first one is to include more physic and dynamic in chemical models. With ALMA in particular, we will have access to a spatial resolution never achieved before. Our view of the chemical distribution of molecules in star forming regions will be determined by small-scale structures. All these improvements of the models however are pointless if the chemistry is wrong. Current chemical networks contain more than 4000 reactions. In addition to the fact these networks are probably incomplete for the most complex molecules, only a small percentage of them have been studied in the ISM temperature range (between 10 and 300 K). The accuracy of chemical models relies on reaction rate coefficients computed and measured by chemists and physicists. Studying a reaction in detail can take years. For this reason, sensitivity analysis have to be conducted to guide experimentalists and theorists on the most important reactions to be studied to improve chemical modelsRead less <