Since the discovery of the first extrasolar planets around two decades ago, more than a thousand of these worlds have been confirmed and characterized. The wide and unexpected diversity of properties shown by these planetary systems suggest the complexity of the planet formation and evolution processes. Apart from providing indications on the formation of the Solar System, these discoveries have opened many others. Step-by-step, we are providing observational hints to answer them. In particular, the Kepler mission has provided an impressive sample of planet candidates of any kind that can be fully characterized thanks to the technique used and the subsequent ground-based follow-up. This full characterization is important in order to analyze their origin and evolution history. In this thesis, we present our contribution to complete the picture of the evolution of planetary systems. We have performed a comprehensive follow-up of the Kepler candidates by making use of ground-based instrumentation at Calar Alto Observatory. Due to the characteristics of the Kepler mission, the detected transits (due to the pass of an object in front of a star) could be due to other blended configurations mimicking a planetary-like transit. Our work has been centered on ruling out these configurations, confirming the planetary-nature of the transiting objects, and analyzing their properties. To that end, we have carried out a two-phases project making use of different datasets and techniques. The two phases consisted on i) obtaining high-spatial resolution images of a large sample of Kepler candidates owing to unveil possible companions and ii) obtaining high-resolution spectroscopy of a smaller carefully selected sub-sample to monitor the radial velocity of the host star and characterize the physical and orbital properties of the planet. In addition, we have analyzed the Kepler light curve looking for modulations induced by the presence of a planetary-mass or substellar object. The results of this follow-up have yielded to the confirmation of five planets in four host stars. Among them, we have found the closest-in planet orbiting a giant star (Kepler-91 b), being the first confirmed planet known to transit one of these evolved stars. Additionally, we confirmed other close-in giant planet around another giant star (Kepler-432 b), the planet having the most grazing transit known to date (Kepler-447 b), and a two-planet system revolving around a young solar-analog (KOI-372). Besides, our high-resolution images of more than 170 planet host candidates have improved the candidacy of tens of planets and have reported close blended companions in around 18% of the sample. In this dissertation we present the observations and analysis that lead to these results and discuss their relevance in the exoplanetary field.